Document not found! Please try again

Hydrothermal Carbonization of Lignocellulosic Biomass

32 downloads 1877 Views 2MB Size Report
Hydrothermal carbonization (HTC) or wet torrefaction is a pretreatment process for ..... secondary wall (S3) may arrange in two or more orientations. Lastly, in ...
University of Nevada, Reno

Hydrothermal Carbonization of Lignocellulosic Biomass

A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Chemical Engineering

By

Mohammad Toufiqur Reza

Dr. Charles J. Coronella/Thesis Advisor

May, 2011

THE GRADUATE SCHOOL

We recommend that the thesis prepared under our supervision by MOHAMMAD TOUFIQUR REZA entitled Hydrothermal Carbonization of Lignocellulosic Biomass be accepted in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE

Charles J. Coronella, Ph.D., Advisor

Victor R. Vasquez, Ph.D., Committee Member

Glenn C. Miller, Ph.D,, Graduate School Representative

Marsha H. Read, Ph. D., Associate Dean, Graduate School

May, 2011

i

Hydrothermal Carbonization of Lignocellulosic Biomass Abstract Hydrothermal carbonization (HTC) or wet torrefaction is a pretreatment process for lignocellulosic biomass where the biomass is treated with hot compressed water. The solid product of HTC is HTC biochar, which is friable, hydrophobic, and increased in mass and energy densification compared to the raw biomass. HTC biochar also is similar regardless of the type of biomass used. A liquid solution of five carbon and six carbon sugars, along with various organic acids and 5-HMF, is also produced from HTC of lignocellulosic biomass. The gaseous phase product consists mostly of CO2. Different types of lignocellulosic biomass were used for HTC under different conditions. Every type of biomass shows a significant decrease in mass yield with a significant increase in the energy densification ratio. Oxygen content decreases with increasing HTC reaction temperature. The oxygen carbon ratio decreases and as a result, HTC biomass has the same characteristics as a low rank coal. To optimize the reaction temperature and reaction time for HTC, reaction kinetics was studied for loblolly pine, which is a lignocellulosic biomass. A special two-chamber reactor was design and built to perform kinetic studies. It is found that hemicelluloses and cellulose degradation follow two parallel first order reactions. Hemicelluloses degrade much faster than cellulose, as the activation energy is lower for hemicelluloses than cellulose. Lignin behaves as an inert in the studied temperature range of 200-2600C and aqueous solubles are generated almost instantaneously in the reaction scheme. After a certain reaction time, the mass yield of HTC becomes steady, but it varies with the reaction temperature. Lignin in the HTC

ii

biochar shows glass transition behavior in the temperature range of 135-1650C. HTC biochar pellets were produced using a hydraulic press operating at 1400C temperature and 1000 MPa. As the lignin content becomes higher at elevated HTC temperatures, the HTC biochar pellets are more durable, abrasion resistant, and mass and energy denser than raw biomass pellets. The higher heating value (HHV) of the pellets is similar to the HHV of HTC biochar. However, the energy density is significantly higher, as the pellets have a higher mass density. HTC biochar pellets have a lower modulus of elasticity and a higher ultimate compressive strength relative to raw biomass. The trend of modulus of elasticity shows it lessens with increasing HTC treatment temperature. The HTC biochar pellets are more hydrophobic than raw biomass pellets. The equilibrium moisture content (EMC) of the HTC biochar pellets are in the same range as HTC biochar, but the pellets take 7-10 days more to reach equilibrium.

iii

Acknowledgments

First of all, I would like to express my sincere gratitude to my advisor Dr. Charles J. Coronella, for his continued guidance, discussion, motivation, and support during my thesis work. It would have been impossible for me to complete this work without his mentoring and moral support. I would like to express my sincere thanks to the committee members, Dr. Victor R. Vasquez, and Dr. Glenn Miller for their valuable comments and suggestions. Special thanks to Ms. Joan G. Lynam, the most helpful person in University of Nevada Reno (UNR), for her enormous invaluable support on my research, discussions, and collaborative studies. She has provided a lot of technical suggestions, personal encouragements, emotional and moral supports. It would have been impossible without her supports. My research group Tapas C. Acharjee, Cody Wagner, Mike Matheus, Jason Hastings, Kevin Schmidt, Cody Niggemeyer, and Chris Moore need special appreciation for their

assistance and creating a joyful environment in the laboratory. Dr. Alan Fuchs research group has been always supportive and accommodating. Thanks to Joko Sustrino and Irawan Paramudiya for their support in analytical measurements. Dr. Qizhen Li is been very generous to me. I used some of her analytical instruments and her valuable advice has been very helpful.

iv

Dr. Subramanyan Ravi’s lab has been always resourceful and his group is always helpful. Special thanks to Dr. Bratindra Mukharjee and York R. Smith for their invaluable support in analytical instrumentation. My parents deserve credit for all my achievements. They have supported me whole heartedly in all my endeavors. Special thanks to Ms. Eriko Mukaibo to support me and stay besides me in all my hard times. She has been my source of inspiration throughout my research. Thanks to my brothers and entire family member for their love and support. Finally, I want to thank the US Department of Energy for their financial support. I gratefully acknowledge meaningful discussions with Larry Felix from the Gas Technology Institute (GTI), Dr. Kent Hoekman from the Desert Research Institute (DRI).

v

CONTENT

ABSTRACT.......................................................................................................................i ACKNOWLEDGMENTS……………………………………………………………...iii FIGURES………..…………………………………………………………….………..ix TABLES……..…………………………………………………………...…………….x ii CHAPTER 1: INTRODUCTION…………….………………………………………….1 1.1 Background................................................................................................1 1.2 Administrative Srtategy………………………….………………………2 1.3 Lignocellulosic Biomass………………....………………………………3 1.3.1 Cell Wall Structure………..………………………………………4 1.3.2 Chemical Structure...……………..………………………....6 1.3.2.1Cellulose…..……………………...………………………..7 1.3.2.2 Hemicelluloses…………………………………………….8 1.3.2.3 Lignin……………………………………………………..10 1.3.2.4 Water Extractives ………………………………………...11 1.4 Conversion Routes for Improved Fuels from Biomass ………….…..…11 1.4.1 Conversion of Dry Biomass……………………………………...12 1.4.2 Conversion of Wet Biomass ………………………………….….13 1.5 Benefits of Pretreatment of Lignocellulosic Biomass…………………...14 1.6 Hydrothermal Carbonization …………………………….……………...15 1.7 Properties of Hot Compressed Water ……………………………….......16

vi

1.8 Project Objectives ……………………………………………………….18 1.9 Organization of Thesis …………………………………………………..19 1.10 References……………..…………….………………………………….20 CHAPTER 2: WET TORREFACTION OF LIGNECELLULOSIC BIOMASS……….27 2.1 Introduction...............................................................................................28 2.2 Experimental Section ……………………………………………………30 2.2.1 Materials………………………..……...………………………….30 2.2.2 Wet Torrefaction………………………………………………….30 2.2.3 Ultimate Analysis…………………………………………………31 2.2.4 Heat of Combustion………………………………………………31 2.2.5 Fiber Analysis ……………………………………………………31 2.2.6 Scanning Electron Microscopy …………………………………..32 2.3 Results and Discussion…………………………………………………..32 2.3.1 Effects of reaction temperature, holding time, water to biomass ratio, and particle size…………………………………………………...32 2.3.2 Change in biomass and temperature effects on mass yield, energy densification ratio, and energy yield…………………………………...34 2.3.3 Incresed Aqueous Solubility and Friablity ……………………....38 2.3.4 Scanning Electron Microscope Images ………………………….39 2.3.5 Recycling of Liquid Product …………………………………….42 2.3.6 Ultimate Analysis ………………………………………………..42 2.4 Conclusion……………………………….………………………………44

vii

2.5 References ………………………………………………………………46 CHAPTER 3: KINETIC STUDY OF HYDROTHERMAL CARBINIZATION OF LIGNOCELLULOSIC BIOMASS……………………………………………………..51 3.1 Introduction..............................................................................................52 3.2 Experimental Section …………………………………………………...54 3.2.1 Biomass ………..……...…………………………………………54 3.2.2 Hydrothermal Carbonization …………………………………….54 3.2.3 Fuel Content…………………………………..….………………58 3.3 Results and Discussions…………………………………........................58 3.3.1 Hydrothermal Carbonization of Loblolly Pine……………..…….58 3.3.2 Kinetic Model for Hydrothermal Carbonization …….…...……...63 3.3.3 Parameter Estimation of Mass Yields Curves ...............................66 3.3.4 Kinetic Parameters of Hydrothermal Carbonization …..………...68 3.4 Conclusion……………………………….……......…..…………………70 3.5 References ………………………………………………………………72 CHAPTER 4: PELLETS FROM PRETREATED BIOMASS…………………………74 4.1 Introduction..............................................................................................76 4.2 Materials and Methods……………………………………………….....80 4.2.1 Biomass and Chemicals……...…………………………………...80 4.2.2 Wet Torrefaction ……………..……………………………….....81 4.2.3 Pelletization Technique...………………………………………...81 4.2.4 Abrasion Index and Durability……..............................................82 4.2.5 Equilibrium Moisture Content……...……………..…….………82

viii

4.2.6 Higher Heating Value………...……….……….…………………83 4.2.7 Digital Scanning Calorimetry………………………….…………83 4.2.8 Mechanical Strength ……….………………………….…………84 4.3 Result and Discussions…...……………………………………………...84 4.3.1 Glass Transition Behavior of HTC Biochar…………………..….84 4.3.2 Mass and Energy Density of Pretreated Biomass Pellets.….…….87 4.3.3 Mechanical Strength of Pretreated Biomass Pellets …..………....90 4.3.3.1 Abrasion Index of Pellets Made from Pretreated Loblolly pine…………………………………………….90 4.3.3.2 Compressive Strength of Pretreated Biomass Pellets…...93 4.3.4 Equilibrium Moisture Content of Pretreated Biomass Pellets…....96 4.4 Conclusions...............................................................................................98 4.5 References.…....………………..…………….…………….……..……100 CHAPTER 5: CONCLUSIONS AND RECOMMENDATIONS FOR FUTURE RESEARCH…………………………...……………………………………..………..105 5.1 Conclusions.............................................................................................105 5.1.1 Wet Torrefaction………...………………………………..…......105 5.1.2 Kinetic Study …………………………..…………………..…...105 5.1.3 Pelletization………………………………………...............…....106 5.2 Recommendations for Future Research ……….………..….…………..107

ix

FIGURES

Figure 1.1: Annual biomass resource potential from forest and agricultural resources…2 Figure 1.2: Typical composition of lignocellulosic biomass…………………………….4 Figure 1.3: Schematic illustration of cell wall (a) and different layers (b) of wood fiber.5 Figure 1.4: Chemical structure of typical lignocellulosic biomass………………………6 Figure 1.5: Schematic illustration of a cellulose chain…………………………………..7 Figure 1.6: Schematic illustration of sugar groups of hemicelluloses……………………8 Figure 1.7: Schematic illustration of partial xylan structure for hardwood (A) and softwood(B)……………………………………………….…………………9 Figure 1.8: Schematic illustration of lignin………………………………….………….10 Figure 1.9: Possible conversion paths for upgrading dry biomass……………………...12 Figure 1.10: Possible conversion paths for upgrading wet biomass…………………….13 Figure 1.11: Physical properties of water with temperature, at 24 MPa………...……...17 Figure 2.1: Effect of temperature on mass yield (a), energy yield (b) for loblolly pine, pelletized corn stover, Tahoe mix (a mixture of Jeffrey pine and white fir from the Tahoe forest), switch grass, and rice hulls………………………..34 Figure 2.2: Effect of temperature on energy densification ratio for loblolly pine, pelletized corn stover, Tahoe mix (a mixture of Jeffrey pine and white fir from the Tahoe forest), switch grass, and rice hulls…………………….35 Figure 2.3: Scanning electron microscope images of (a) raw loblolly pine and

x

loblolly pine pretreated at (b) 200˚C, (c) 230˚C, and (d) 260˚C with the magnification of 250 times. (e) Possible arrangement of chemical components in lignocellulosic biomass……………………………………40 Figure 2.4 Scanning electron microscope images of (a) raw loblolly pine and loblolly pine pretreated at (b) 200˚C, (c) 230˚C, and (d) 260˚C with 500 times magnification………………………………………………………...….…41 Figure 2.5: Van Krevelen diagram of loblolly pine, rice hulls, and corn stover showing effect of 275˚C wet torrefaction pretreatment……………………44 Figure 3.1 Two chamber reactor (a), schematic of the two-chamber reactor system (b), the components: 1. Bottom chamber; 2. Top chamber; 3. Ball valve; 4. Pressure relief valve; 5. Water-cooling coil; 6. Radiant heater; 7. Temperature indicator; 8. PID temperature controller…………..………55 Figure 3.2: Making of cigarette-shape capsule sample holder………………………….56 Figure 3.3: Method for placing biomass in hot compressed water for rapid sample heating. (Left to right) (a) water is heated while sample in top chamber, (b) open the ball valve allowing sample to drop into the bottom chamber, (c) sample is immersed in the hot water and ball valve is closed..57 Figure 3.4: Mass yield of hydrothermal carbonized loblolly pine at reaction temperatures. Curves only indicate the tendency of the change of mass yield………………………………………………………………………....61 Figure 3.5: Fuel content of hydrothermal carbonized loblolly pine (on a dry basis) at reaction temperatures. The higher heating value (HHV) of raw loblolly pine is 19.22 MJ·kg-1. Curves only indicate the tendency

xi

of the change of fuel content…………………….....………………………62 Figure 3.6: Model prediction of hydrothermal carbonization of loblolly pine for long reaction time sat reaction temperatures………………………..….67 Figure 3.7: Kinetic modeling of hydrothermal carbonization of loblolly pine at reaction temperatures…………………………………………………….68 Figure 3.8: Determination of kinetic parameters of hydrothermal carbonization of loblolly pine……………………………………………………………...69 Figure 4.1: Deformation mechanism of powder particle under pressure………………..77 Figure 4.2: DSC curves (slope of heat flow versus temperature) for determination of glass transition temperature of extracted lignin from HTC biochar and raw loblolly pine…………………………….85 Figure 4.3: Mass density of the pellets of raw loblolly pine and HTC biochar pretreated at different temperatures………………………….88 Figure 4.4: Abrasion index of the pellets of raw loblolly pine and HTC biochar pretreated at different temperatures………………………….92 Figure 4.5: Determination of modulus of elasticity of HTC biochar pellets…………….95

xii

TABLES

Table 2.1: Effect of temperature, holding time, water biomass ratio, and particle size on mass yield, energy densification, and energy yield…….33 Table 2.2a. Fiber analysis of raw switch grass, corn stover, Jeffery pine, rice hulls, and loblolly pine………………………………………………….36 Table 2.2b. Fiber analysis of loblolly pine pretreated at 200˚C and 260˚C……………..37 Table 2.3: Ultimate analysis of different biomass both raw and pretreated condition….43 Table 3.1. Hydrothermal carbonization of loblolly pine. The high heating value (HHV) of raw loblolly pine is 19.22 MJ·kg-1………………………………..60 Table 3.2. Fiber Analysis of raw loblolly pine and predicted reaction type of different constituents in loblolly pine………………………………………………….64 Table 3.3. Kinetic parameters of hydrothermal carbonization of loblolly pine…………………………………………………………………70 Table 4.1: Mass and energy density of loblolly pine wood, pellets of raw loblolly pine, pellets pretreated at 2000C (HTC-200), 2300C (HTC-230), and 2600C (HTC-260)……………...89 Table 4.2: Abrasion index and durability of pellets of loblolly pine and HTC biochar pretreated at different temperatures…………………..……….91 Table 4.3 Ultimate compressive strength and modulus of elasticity of HTC biochar pellets………………………………………………………….94 Table 4.4: EMC of the HTC biochar pellets at different relative humidities……………97

1

Chapter 1

Introduction

1.1 Background Using lignocellulosic biomass for producing energy is not a new concept. Since the eighteenth century, humans have adopted various ways converting biomass into food, cloth, and household items[1]. The transportation and handling of solid biomass is expensive considering its low fuel value with respect to fossil fuels. The discovery and use of crude oil developed the modern transportation systems. Fossil fuel, which is fossilized solar radiation energy captured by plants in past eons, provides about 80% of the energy used today[2]. It is not considered renewable or sustainable due to its finite reserves and environmental difficulties from emissions of greenhouse gases, mainly CO2[4]. Unstable market prices, limit availability, and emissions has caused us to look into renewable and alternative energy sources[5]. Lignocellulosic biomass is an alternative energy resource. It is one of the largest compared to coal, oil, and natural gas. It is abundant in most areas, but it has low carbon content, it is easily biodegradable (hydrophilic), and it has a high handling costs. However, it is an attractive feedstock as a renewable energy source and its relatively short carbon cycle. Its use also decreases the amount of greenhouse gases to the environment as biomass takes CO2 from the environment while growing and releases less NOx and SOx in thermochemical conversion compared to fossil fuels[3]. The price of the lignocellulosic biomass such as rice hulls, switch grass, softwoods, hardwoods etc. are relatively stable and less than the fluctuating fossil fuel price. However, lignocellulosic biomass should be converted to a more

2

hydrophobic, homogenized, energy dense fuel, with better handling characteristics from its raw conditions. Because of the low carbon content and hydrophilic behavior, it is not favorable to use for energy or transportation fuel directly. 1.2 Administrative Strategy The United States is an importer of petroleum and derived products. In 2008, the US

domestic daily oil production was 8.5 million barrels with a daily demand of 19.5 million barrels[6]. In 2006, the US had 2% of world oil reserves and used 24% of the world supply[7]. Nearly 60% of the oil used in the US is imported to meet the current energy demand. To encourage domestic production of renewable fuels, the Energy Independence

and Security Act (EISA) of 2007 clearly required that at least 36 billion gallons of renewable fuel must be produced and used in US by 2022[8-10].

Figure 1.1 Annual biomass resource potential from forest and agricultural resources[11]. In recent years, a numerous research studies and projects are developing to abundantly available biomass resources of the country for biofuels applications. The US

3

government also encourages these development. The billion ton vision report concludes that land resources in the US are capable of producing a sustainable supply of biomass sufficient to displace 30% or more of current petroleum consumption with biofuels by the year 2030[11]. Figure 1.1 shows that more than a billion tons of biomass per year from forest and agricultural resources is available for sustainable biofuels production. Due to its fibrous nature (low density, and low heating value) raw lignocellulosic biomass is not considered an ideal fuel. It requires some modification to make it an attractive competitor with fossil fuel[13]. Currently, production of ethanol by fermentation of sugars (from corn grain or sugar cane) and the transesterification of fatty acids from soy, canola, and other natural oils to biodiesel are the two major chemical pathways to produce biofuels commercially. In both cases, human food sources are used to make the biofuel, which is a controversial because of the increasing demand for food in the world[14]. Lignocellulosic biomass has the potential to play an important role in the energy market. Nevertheless, it requires more research and pilot-plant studies before successful commercialization. It can be a good replacement of coal or blended with coal for cofiring. Ethanol production from lignocellulosic biomass by fermenting sugars, is in pilotplant stage[15]. The feasibility of higher hydrocarbons (C16-C18) such as diesel production from fatty acid (derived from biomass) is considering too. 1.3 Lignocellulosic Biomass Most plants have four main constituents, namely, cellulose, hemicelluloses, lignin, and water extractives. The composition varies among plant types. Most of the softwoods contain

4

more cellulose than grasses type biomass[17]. Figure 1.2 shows the main constituents and their composition ranges among various lignocellulosic biomass.

Figure 1.2 Typical composition of lignocellulosic biomass[16]. 1.3.1 Cell Wall Structure The basic cell structure is the same for all lignocellulosic biomass, but the thickness and composition can be different depending on the type of biomass. The basic model of the wood cell wall structure is well understood[18-20]. Figure 1.3(a) shows the basic structure of the wood cell wall, while Figure 1.3(b) shows the relative thickness of the layers of the cell wall for typical wood. Middle lamella (MT) is a kind of glue type component which can be found in the cells gap and it is the one which binds the cells to each other. Towards the inside, the cell wall is called the primary wall (P). The outer and

5

inner surfaces are two main constituents of the primary wall. Following the primary wall is the secondary wall, which consists of three layers. They are the outer layer (S1), middle layer (S2), and inner layer (S3). In the outer layer of the secondary wall (S1), the microfibrils are oriented in a cross-helical structure (S helix and Z helix), while the middle layer of the secondary wall (S2), which is the thickest layer, has relatively consistent orientation of microfibrils. In contrast, the microfibrils of the inner layer of the secondary wall (S3) may arrange in two or more orientations. Lastly, in some cases, there is a warty layer (W) on the inner surface of the cell wall. In addition, some authors mention that there is a tertiary wall (T) between S3 and W.[22]

()() (a)

(b)

Figure 1.3: Schematic illustration of cell wall (a) and different layers [18] (b) of wood fiber[21].

6

1.3.2 Chemical Structure Figure 1.4 shows the typical chemical structurte of the lignocellulosic biomass. The chemical components of lignocellulose can be divided into four major components. They are cellulose, hemicelluloses, lignin, and extractives.

Figure 1.4: Chemical structure of typical lignocellulosic biomass [23]. Generally, the first three components have high molecular weights and contribute much mass, while the latter component is of small molecular size, and it is available in small quantities. Based on weight percentage, cellulose and hemicelluloses are higher in

7

hardwoods compared to softwoods and wheat straw, while softwoods have higher lignin content. In general, cellulose can be found as a bundle in the raw lignocellulosic biomass, while hemicellulose is the spiral cover of the cellulose and lignin separates the bundles from each other like a stratum (Fig. 1.4). Aqueous solubles can be found as a thin layer over the whole configuration[24].

1.3.2.1 Cellulose The cellulose content of wood varies between species in the range of 38-55 %[17]. Cellulose is a linear polymer chain which is formed by joining the anhydroglucose units into glucan chains. These anhydroglucose units are bound together by β-(1,4)-glycosidic linkages. Due to this linkage, cellobiose is established as the repeat unit for cellulose chains (Figure 1.5). The degree of polymerization (DP) of native cellulose is in the range of 10,000-15,000.[25]

Figure 1.5: Schematic illustration of a cellulose chain[19]. By forming intramolecular and intermolecular hydrogen bonds between OH groups within the same cellulose chain and the surrounding cellulose chains, the chains tend to

8

arrange in parallel and form a crystalline supermolecular structure. Then, bundles of linear cellulose chains (in the longitudinal direction) form a microfibril which is oriented in the cell wall structure[22]. 1.3.2.2 Hemicelluloses Unlike cellulose, hemicelluloses consist of different monosacharide units. In addition, the polymer chains of hemicelluloses have short branches and are amorphous. Because of the amorphous morphology, hemicelluloses are partially soluble or soluble in water. The backbone of the chains of hemicelluloses can be a homopolymer (generally consisting of single sugar repeat unit) or a heteropolymer (mixture of different sugars). Formulas of the sugar component of hemicelluloses are listed in Figure 5.

Figure 1.6: Schematic illustration of sugar groups of hemicelluloses[17].

9

Among the most important sugars of the hemicelluloses component is xylose. In hardwood xylan, the backbone chain consists of xylose units which are linked by β-(1,4)glycosidic bonds and branched by α-(1,2)-glycosidic bonds with 4-O-methylglucuronic acid groups. In addition, O-acetyl groups sometime replace the OH groups in position C2 and C3 (Figure 1.7 A). For softwood xylan, the acetyl groups are fewer in the backbone chain. However, softwood xylan has additional branches consisting of arabinofuranose units linked by α-(1,3)-glycosidic bonds to the backbone (Figure 1.7 B)[22].

Figure 1.7: Schematic illustration of partial xylan structure for hardwood (A) and softwood (B)[17]

10

1.3.2.3 Lignin

Lignin is a complex, crosslinked polymer that forms a large molecular structure. Lignin gives mechanical strength to wood by gluing the fibers together (reinforcing agent) between the cell walls. It is often associated with the cellulose and hemicellulose to make lignocellulosic biomass. Softwood lignins are formed from coniferyl alcohol[24].

Figure 1.8: Schematic illustration of lignin[26]. Hardwood lignins have both coniferyl and sinapyl alcohol as monomer units. Grass lignin contains coniferyl, and sinapyl. Lignin also serves as a disposal mechanism for metabolic waste[21]. The monomeric building units of lignin are shown in Figure 1.8.

11

The guaiacyl unit is dominant in the softwoods[27]. In contrast, syringyl units are dominant in hardwoods[28].

1.3.2.4 Water Extractives

Extractives are the organic substances which have low molecular weight and are soluble in neutral solvents. Resin (combination of the following components: terpenes, lignans and other aromatics), fats, waxes, fatty acids and alcohols, terpentines, tannins and flavonoids are categorized as extractives. They only represent between 4-10 % of the total weight of dry wood, and the contents of extractives vary among wood species, geographical site, and season. The extractives can be found mostly in resin canal and ray parenchyma cells, with small amounts in middle lamella and cell walls of tracheids. Some extractives are toxic and this is an advantage for the wood in resisting attack by fungi and termites[22].

1.4 Conversion Routes for Improved Fuels from Biomass Conversion processes are available or under development for both wet and dry feedstocks. Examples of wet biomass are: sewage sludge, sugar solutions, algae suspensions, and waste animal manure from biomass processing or from biorefineries. Biomass with moisture less than 30 wt% is classified as dry biomass[13]. Examples of dry biomass are: wood, straw, or other sun dried waste. Of course wet biomass can be dried with energy from other sources, but this is not always the most efficient or economical way to operate.

12

1.4.1 Conversion of Dry Biomass Mechanical treatment and compacting could be used efficiently in close proximity to the production sites. For example, pressing the oil from oil rich seeds is typically near to the farm of origin. For dry biomass, apart from combustion, fast or slow pyrolysis can be applied to produce an oil like substance, char, and gas. Also, gasification to fuel gas or to syngas for production of synthetic fuel is a possible route. Moreover, solvolysis using organic solvents can be applied[29,30]. Different combinations of pretreatment, intermediate conversion, and final conversion can be used depending on local options and/or economics, as indicated in Figure 1.9.

Figure 1.9: Possible conversion paths for upgrading dry biomass[13].

13

In a fast pyrolysis process biomass is very quickly heated to approximately 500oC and the vapors are rapidly quenched to produce a liquid (up to 70 wt% of the biomass)[31]. which, after stabilization, can be stored and transported, for further upgrading. The liquid product still contains a large amount of oxygen (± 40-50 %)[32]. Fuel gas and char are produced as byproducts, part of which can be used to energize the process. 1.4.2 Conversion of Wet Biomass Wet biomass (see Figure 1.10) can be converted into improved fuels via biological routes, such as anaerobic digestion to methane rich gas or fermentation to alcohols. Conventionally these routes are limited to certain carbohydrate fractions of biomass. However, for the so-called second generation conversion processes, enzymes and pretreatment options are being developed that target the lignocelulosic biomass in a broader sense[32].

Figure 1.10: Possible conversion paths for upgrading wet biomass[13].

14

For wet biomass conversion, processes which do not require water evaporation are desired, because, the water evaporation requires additional energy. In addition to biological conversion, conversion in hot compressed water, both sub- and super-critical, is possible to produce hydrophobic liquids, solids, and gasses[33,34]. Apart from hot compressed water, other solvents have been used for biomass conversion[35-39]. However, this is not a topic of the present work. By combining dry and wet conversion routes, a wide spectrum of interconnected thermochemical biomass conversion routes toward final products is possible, and may be used in a biorefinery. In such complex concepts, hydrothermal conversion can be applied to produce intermediate energy carriers, in primary conversion steps like gasification and deoxygenation; or it can be used for working up of side/ waste streams from conversion processes of biomass to food, feed, or chemicals[13]. 1.5 Benefits of Pretreatment of Lignocellulosic Biomass Lignocellulosic biomass is one of the most promising fuel sources in the world. It has been used for centuries to heat or make power by direct firing. But with some conversion it can produce solid, liquid, or gas fuels. It is relatively cheap compared to fossil fuels and the process is renewable, sustainable, and environmentally friendly. Unfortunately, diverse biomass feedstocks exhibit diverse handling characteristics, complicating their usage. This challenge is further compounded by the expensive logistics of seasonal availability in the case of agricultural wastes or wide distribution in the case of forestry. These difficulties lead to the necessity of pretreatment techniques[40]. They may be thermal, biological, or hydrothermal in nature[41]. Each of them has advantages and disadvantages, but improve the energy utilization of biomass.

15

Selecting the proper pretreatment depends on the type of biomass and type of output. The following advantages can be found for a given pretreatment[14]. (1) Low cost of chemicals for pretreatment, neutralization, and subsequent conditioning (2) Minimal waste production (3) Limited size reduction because biomass milling is energy-intensive and expensive (4) Fast reactions and noncorrosive chemicals to minimize pretreatment reactor cost (5) Pretreatment should facilitate recovery of lignin and other constituents for conversion to valuable co-products and to simplify downstream processing. For those various benefits, it creates a large scope of pretreatment of lignocellulosic biomass. 1.6 Hydrothermal Carbonization Hydrothermal carbonization (HTC), also known as hydrothermal pretreatment or wet torrefaction, is a thermo-chemical conversion technique which uses liquid subcritical water as a reaction medium for conversion of wet biomass and waste streams[40]. It is usually performed at temperatures higher than 180 °C, at pressures high enough to ensure liquid water, and under an inert atmosphere. Reaction time has been reported to be 1 minute to several hours, although most of the reaction seems to occur within the first 20 minutes[42,43]. Additives, such as acids or bases, can affect the products formed. Undried biomass and water may be used in this process. As both reactant and solvent, water shows different physical and chemical properties depending on the operating conditions[44]. At temperatures between 227 and 327 °C, water may act as both a base and an acid because its ionic product is maximized. In addition, water’s dielectric

16

constant is decreased at these temperatures so that it acts more like a non-polar solvent[45]. During hydrothermal pretreatment, hemicelluloses and cellulose are hydrolyzed to oligomers and monomers,[46,47] while lignin is mostly unaffected. The solid product, also known as biochar, has reduced equilibrium moisture content, so it is less likely to rot in storage[48]. The pretreated solid is quite friable and might be made into pellets which can be fed to a gasifier or coal power plant easily. The liquid products can also be further fractionated by means of extraction with polar organic solvent(s)[4953]. The solvent-soluble fraction is then the desired product, part of which can be upgraded to transportation fuel quality by catalytic hydro-deoxygenation[54-55]. The production of an intermediate suitable for refining and upgrading into transportation fuel is the preferred aim of HTC. Therefore this option is intensively studied, with the focus on minimizing the yields of the reaction byproducts and on the product separation. Hydrothermal carbonization performance is better than the dry torrefaction at the same temperature of pretreatment. Yan et al. reported that the HTC biochar pretreated at 2600C has higher mass and energy density than the dry torrefied biochar treated at 3000C[40]. The mass and energy balance of the hydrothermal carbonization was calculated and reported[60]. The effect of adding different acids and salts were verified[61]. 1.7 Properties of Hot Compressed Water From the phase diagram of water, the critical point is at 374°C and 22.1 MPa. Liquid water, below the critical point, is subcritical and above is supercritical. Water shows a good solubility with different compounds due to its dielectric point and density even at ambient condition (25°C and 0.1 MPa).

17

With the introduction of heat, the H-bond of water starts weakening, allowing the dissociation of water into acidic hydronium ions (H3O+) and basic hydroxide ions (OH−). The structure of every substance changes significantly near the critical point. In water, the infinite network of hydrogen bonds is broken near the critical point and water then exists as separate clusters with a chain structure[56]. In fact, dielectric constant of water decreases considerably near the critical point, which causes a change in the dynamic viscosity and also increases the self-diffusion coefficient of water[57]. Supercritical water has liquid-like density and gas-like transport properties, and behaves very differently than it does at room temperature. The ionic product of water reaches its maximum value at temperatures between 227 °C and 327 °C depending upon the pressure. In this temperature range, the ionic product is greater by 1 or 2 orders of magnitude than at ambient temperature[58].

Figure 1.11: Physical properties of water with temperature, at 24 MPa[59].

18

The thermophysical properties of water, such as viscosity, ionic product, density, and heat capacity, also change dramatically in the supercritical region with only a small change in the temperature or pressure, resulting in a substantial increase in the rates of chemical reactions. It is interesting to note that the dielectric behavior of 200°C water is similar to that of ambient methanol, 300°C water is similar to ambient acetone, 370°C water is similar to methylene chloride, and 500°C water is similar to ambient hexane[59]. 1.8 Project Objectives The main objective of this thesis is to understand and optimize the hydrothermal carbonization (HTC) of lignocellulosic biomass. This includes the finding of important variables and their effect on this process. The effect of variables in the HTC is led the process verification. Characterization of the solid output is another objective. The effect of HTC applied to different types of lignocellulosic biomass is conducted, which leads to the process applicability for every kind of lignocellulosic biomass. Lignocellulosic biomass is usually diverse and if only one process is applicable for all type of biomass, then it will be very attractive. Kinetics of the procedure is investigated using one of the common lignocellulosic biomass which could be applied to find the optimum conditions for HTC process. This will facilitate the process design and reaction condition. The HTC biochar is friable, hydrophobic and densefied (both mass and energy). But the solid handling is not solved by HTC. Pelletization can be a promising factor of making the HTC bochar denser and it can be easy to transport and use. The pellets are usually less dusty, so, the mass loss during the transport will be less. Lignin is a cross linked polymer and it shows the glass transition behavior around 1400C. It can be a potential binder for the pelletization of HTC biochar. So, the final objective of this study is to make pellets

19

from the HTC biochar by applying the glass transition condition of lignin. For that objective, the glass transition behavior of lignin is verified and the effect of lignin in pelletization of HTC biochar is examined. 1.9 Organization of Thesis Chapter 2 shows the effect of the HTC process on different types of lignocellulosic biomass. The characterization of its solid product is another part of this chapter. In this chapter, the reaction time was constant at 5 minutes and different biomass were used for the HTC process. Chapter 3 explains the kinetic study of the HTC process for loblolly pine, one of the common lignocellulosic biomass. A specially-designed two-chamber reactor was utilized to perform hydrothermal carbonization of lignocellulosic biomass isothermally. The hydrothermal carbonization of loblolly pine was performed at various reaction times for temperatures ranging from 200 to 260oC. The results show that hydrothermal carbonization can be achieved in a short period of time and that two parallel first-order reactions, namely for conversion of hemicelluloses and cellulose, can characterize the HTC process. In chapter 4, the process of making pellets from HTC biochar is introduced. Pretreated loblolly pine is used here as lignocellulosic biomass. Different conditions for pelletization are discussed in this chapter. It is found that lignin can act as a natural binder for this process which validates the kinetic study which shows that lignin is an inert in this temperature range. In chapter 5 summarizes some conclusions drawn from the previous chapters of this thesis and contains recommendations for further research.

20

1.10 References [1] Demirbas, M. F. (2006). Current Technologies for Biomass Conversion into Chemicals and Fuels, Energy Sources, Part A, 28, 1181–1188. [2] IEA: Key world energy statistics: 2008 [3] Pastircakova, K., Determination of trace metal concentrations in ashes from various biomass materials. Energy Edu. Sci. Technol. 2004, 13, 97-104. [4] Yu, Y.; Lou, X.; & Wu, H. (2008). Some Recent Advances in Hydrolysis of Biomass in Hot-Compressed Water and Its Comparisons with Other Hydrolysis Methods, Energy & Fuels, 22, 46–60 [5] Huber, G. W.; Iborra, S., & Corma, A. (2006). Synthesis of Transportation Fuels from Biomass: Chemistry, Catalysts, and Engineering, Chemical Reviews, 106, 4044-4098. [6] International Energy Statistics, EIA, US Energy Information Administration, US Department of Energy. [7] http://www-cta.ornl.gov/data/chapter1.shtml. (Feb, 2010) [8] Biomass multiyear program. Office of the Biomass Program, Energy Efficiency and Renewable Energy, U.S. Department of Energy, Washington, D.C. 2008. [9] Huber, G.W.; Dale B.E. Sci. Am. 2009, July, 52. [10] Biofuels create green jobs: Growing transportation fuels and the nation’s economy. Office of the Biomass Program, Energy Efficiency and Renewable Energy, U.S. Department of Energy, Washington, D.C., 2008.

21

[11] Perlack, R. D.; Wright, L. L.; Turhollow, A. F.; Graham, R. L.; Stokes, B. J.; Erbach, D. C. 2003. Biomass as feedstock for a bioenergy and bioproducts induatry: The technical feasibility of a billion-ton annual supply; US Department of Energy [12] Ragauskas A. J.; Williams C.K.;Davison B.H.; Britovsek G.; Cairney J.; Eckert C.A.; Frederick Jr W.J.; Hallett J.P.; Leak D.J.; Liotta C.L.; Mielenz J.R.; Murphy R.; Templer R. & Tschaplinski T.(2006). The Path Forward for Biofuels and Biomaterials, Science, 311, 484. [13] Knežević, D. (2009). Hydrothermal Conversion Biomass, MS thesis, University of Twente. [14] Huber, G. W.; Iborra, S., & Corma, A. (2006). Synthesis of Transportation Fuels from Biomass: Chemistry, Catalysts, and Engineering, Chemical Reviews, 106, 40444098. [15] Piccolo, C.; Bezzo, F. (2009). A techno-economic comparison between two technologies for bioethanol production from lignocellulose, Biomass and Bioenergy, 33, 478-491. [16] http://www.nrel.gov. (Feb, 2010). [17] Kumar, S. (2010). Hydrothermal Treatment for Biofuels: Lignocellulosic Biomass to Bioethanol, Biocrude, and Biochar, PhD Dissertaion, Auburn University. [18] Cote, W. A., Wood Sci. Technol., 15, 1-29 (1981). [19] Fengel, D., and G. Wegener, Wood: Chemistry, Ultrastructure, Reactions, Walter de Gruyter, Berlin, (1989).

22

[20] Moniruzzaman, M., Bioresource Technol., 55, 111-117, (1996). [21] Wardrop, A. B., Occurrence and Formation in Plants, in “Lignin - Occurrence, Formation, Structure and Reactions”, Sarkanen, K. V., and C. H. Ludwig, Eds., Wiley Interprice, N. Y., 19-32 (1971). [22] Ibrahim, M. Clean fraction of Biomass- Steam Explosion and Extraction. MS Thesis, Virginia Tech University, (1998). [23] An overview of Science; July 2009. Boenergy Resource Center. US Dept of Energy. Also available on http://jbei.lbl.gov/assets/docs/brcbrochure_hq.pdf. [24] Acharjee, T. C. Thermal Pretreatment Options for Lignocellulosic Biomass. MS Thesis. University of Nevada. (2010). [25] O' Sullivan, A.C. (1997). Cellulose: the structure slowly unravels, Cellulose, 4,173207. [26] Mohan, D; Pittman, Jr. C.U.; & Steele P.H. (2006).Pyrolysis of wood /biomass for bio-oil: a critical review, Energy Fuels, 20 (3), 848–889. [27] Glasser, W. G., Lignin, in “Pulp and Paper: Chemistry and Chemical Technology”, Casey, J. P., Ed., 3rd. Ed., Vol. I, John Willey & Sons, 39-111, (1980). [28] Martin, R. S., C. Perez and R. Briones, Bioresource Technol., 53, 217-223, (1995). [29] Chornet, E.; Overend, R.P. Biomass liquefaction: An overview. In: Fundamentals of Thermochemical Biomass Conversion (edited by Overend, R.P.; Milne, T .A.; Mudge, L.K.), Elsevier Applied Science, 1985, pp.967. [30] Behrendt, F.; Neubauer, Y.; Oevermann, M.; Wilmes, B.; Zobe, N. Direct Liquefaction of Biomass (review) Chemical Engineering & Technology 2008, 31, 667.

23

[31] Bridgwater, A.V. Principles and practice of biomass fast pyrolysis processes for liquids. Journal of Analytical and Applied Pyrolysis 1999, 51,3. [32] Iogen: a leader in cellulose ethanol Bio-Prospects 2005 Volume 2, 1, Published by Ag-West Bio Inc. [33] Goudriaan, F.; Peferoen, D. G. R. Liquid fuels form biomass via a hydrothermal process. Chem.Eng. Sci. 1990, 45, 2729. [34] Elliott, D. C.; Silva, L. J. The TEES process cleans waste and produces energy. Presented at the R1995 Conference, Geneva, Switzerland, 1-3 Feb. 1995. [35] Wilhelm, D. J; Kam, A.Y.; Stallings, J. W. Transportation fuel from biomass by direct liquefaction and hydrotreating. Symposium papers: Energy from biomass and wastes V, January 26-30, 1981, Lake Buena Vista, Florida, USA, pp.651. [36] Vasilakos, N.P.; Austgen, D.M. Hydrogen-Donor Solvents in Biomass Liquefaction. Ind. Eng.Chem. Proc. Res. Dev 1985, 24, 304. [37] Heitz, M.; Brown, A.; Chornet, E. Solvent Effects On Liquefaction: Solubilization Profiles of a Canadian Prototype Wood, Populus deltoids, in the presence of different solvents. Can. J.Chem.Eng. 1994, 72, 1021. [38] Rezzoug, S. A.; Capart, R. Solvolysis and hydrotreatment of wood to provide fuel. Biomass and Bioenergy 1996, 11, 343. [39] Mun, S. P.; Hassan, E. B. M. Liquefaction of Lignocellulosic Biomass with Dioxane/Polar Solvent Mixtures in the Presence of an Acid Catalyst. J. Ind. Eng. Chem. 2004, 10, 722.

24

[40] Yan, W.; Acharjee, T.C.; Coronella, C.J.; Vasquez, V.R. 2009. Thermal pretreatment of lignocellulosic biomass, Environmental Progress & Sustainable Energy, 28, 435-440. [41] Huber, G. W.; Iborra, S., & Corma, A. (2006). Synthesis of Transportation Fuels from Biomass: Chemistry, Catalysts, and Engineering, Chemical Reviews, 106, 40444098. [42] Funke, A.; Ziegler, F. Hydrothermal Carbonization of Biomass: A Literature Survey Focusing on Its Technical Application and Prospects, 2009.17th European Biomass Conference and Exhibition, Hamburg, Germany. [43] Lu, X., Yamauchi, K., Phaiboonsilpa, N., 2009. Two-step hydrolysis of Japanese beech as treated by semi-flow hot-compressed water. J. Wood Sci. 55, 367-375. [44] Ando, H., Sakari, T., Kobusho, T., Shibata, M., Uemura, Y. Hatate, Y., 2000. Decomposition behavior of plant biomass in hot-compressed water. Ind. Eng. Chem. Res. 39, 3688-3693. [45] Yu, Y., Lou, X., Wu, H., 2007. Some recent advances in hydrolysis of biomass in hot-compressed water and its comparisons with other hydrolysis methods. Energy Fuels 22, 46-60. [46] Petersen, M. O., Larsen, J., Thomsen, M. H., 2009. Optimization of hydrothermal pretreatment of wheat straw for production of bioethanol at low water consumption without addition of chemicals. Biomass Bioenergy 33, 834-840.

25

[47] Funke, A., Ziegler, F., 2010. Hydrothermal carbonization of biomass: A summary and discussion of chemical mechanisms for process engineering. Biofuels Bioproducts & Biorefining-Bioref. 4, 160-177. [48] Acharjee, T. C., Coronella, C. J., Vasquez, V. R., 2011. Effect of thermal pretreatment on equilibrium moisture content of lignocellulosic biomass. Bioresource Tech. 102, 4849-4854. [49] Yokoyama, S-Y; Ogi, T; Koguchi, K; Nakamura, E. Direct liquefaction of wood by catalyst and water. Petroleum Science and Technology 1984, 2,155. [50] Ogi, T.; Yokoyama, S-Y.; Kuguchi K. Direct liquefaction of wood by catalyst part I. Effect of pressure, temperature, holding time and wood/catalyst/water ratio on oil yield, Sekiyu Gakkaishi 1985 28, 239. [51] Krochta, J.M.; Hudson, J.S.; Drake, C.W.; Mon, T.R.; Pavlath, A.E. Thermal Degradation Of Cellulose in Alkali. In: Fundamentals of Thermochemical Biomass Conversion: An internationa Conference, Estes Park Colorado, 1985, pp. 1073. [52] Ogi, T.; Minowa, T.; Dote, Y.; Yokoyama, S-Y. Characterization of oil produced by the direct liquefaction of Japanese oak in an aqueous 2-propanol solvent system. Biomass and Bioenergy 1994, 7, 193. [53] Cheng, L.; Ye, P., X; He,R.; Liu, Sh. Investigation of rapid conversion of switchgrass in subcritical water. Fuel processing technology 2009, 90, 301. [54] Baker, E.G. ; Elliott, D.C. Catalytic hydrotreating of biomass-derived oils. In J. Soltes & T.A. Milne (eds.), Pyrolysis oils from biomass, ACS Symp. Series 376 (Washington, DC: American Chemical Society, 1988).

26

[55] Moffatt, J. M.; Overend, R. P. Direct liquefaction of wood through solvolysis and catalytic

hydrodeoxygenation: an engineering assessment, Biomass 1985, 7, 99.

[56] Kalinichev, A. G.; Churakov, S. V., Size and topology of molecular clusters in supercritical water: a molecular dynamics simulation. Chem. Phys. Letters 1999, 302, 411-417. [57] Marcus, Y., On transport properties of hot liquid and supercritical water and their relationship to the hydrogen bonding. Fluid Phase Equilibr. 1999, 164, 131-142. [58] International Energy Agency (IEA). (2008). World Energy Outlook. [59] Kritzer, P.; Dinjus, E., An assessment of supercritical water oxidation (SWO). Existing problems, possible solutions and new reactor concepts. Chemical Engineering Journal (Amsterdam, Netherlands) 2001, 83, (3), 207-214. [60] Yan, W., Hastings J.T., Acharjee, T.C., Coronella, C.J., Vasquez, V.R., 2010. Mass and energy balance of wet torrefaction of lignocellulosic biomass. Energy Fuels 24, 4738-4742. [61] Lynam, J. G., Coronella, C. J., Yan, W., Reza, M. T., and Vasquez, V. R., 2011. Acetic acid and lithium chloride effects on hydrothermal carbonization of lignocellulosic biomass. Bioresour. Technol. doi:10.1016/j.biortech.2011.02.035.

27

Chapter 2

Wet torrefaction of lignocellulosic biomass

Wet torrefaction is a process which pretreats lignocellulosic biomass for subsequent thermochemical conversion. Biomass reacts in hot compressed water at temperatures between 200 ˚C and 275 ˚C at pressures required to maintain liquid water for times less than 30 min. Increasing reaction temperature, the only process variable with a significant effect, increases energy density and decreases mass yield. Ultimate analysis indicates that the pretreated solid product has significantly reduced atomic oxygen and increased carbon content. Thus, from lignocellulosic biomass, wet torrefaction can produce a solid product similar to low rank coal.

28

2.1 Introduction As an indefinitely renewable resource, biomass has the potential to reduce the world’s dependence on fossil fuels. Fossil fuels are by their nature limited, as well as mostly found in areas that are politically unstable. In addition, the burning of fossil fuels increases the amount of CO2 in our atmosphere, contributing to global warming. As a renewable energy source, biomass uses as much CO2 in its growth as it releases as a fuel. Thus, it is considered to be carbon neutral, and not to contribute to global warming. In the United States alone, over 367 million dry tons of biomass per year is estimated available from forested areas, while 998 million dry tons of biomass per year could be supplied by agricultural lands[1]. Although ethanol can be made from corn and other renewable food crops, many believe it unethical to use food for liquid transportation fuels. Lignocellulosic biomass is not only a non-food resource, but can even be a waste product, as in the case of rice hulls or corn stover. Lignocellulosic biomass can be used directly as fuel, such as logs used in a fireplace. After mechanical size reduction, it also can be used for co-firing in a coal power plant[2]. In addition, liquid biofuels can be produced from low-cost lignocellulosic biomass by either biochemical conversion routes or thermochemical conversion routes[3]. Unfortunately, diverse biomass feedstocks exhibit diverse handling characteristics, complicating their usage. This challenge is further compounded by the expensive logistics of seasonal availability in the case of agricultural wastes or wide distribution in the case of forestry. Thus, there is a need to develop pretreatment processes that produce a solid fuel from lignocellulosic biomass with increased higher heating value (HHV),

29

increased hydrophobicity for better storage, less mass to transport, and homogeneous handling characteristics. Wet or dry torrefaction can be used for this purpose[4]. Dry torrefaction, also called mild pyrolysis, typically is performed in an inert atmosphere at 0.1 MPa between 225°C and 300°C using reaction times which vary between 30 min and several hours[4-6]. Dry torrefaction produces a reduced higher heating value (HHV) solid product compared to wet torrefaction when performed at similar temperatures and for similar reaction times[7]. Another disadvantage of dry torrefaction is the energy needed to dry the feedstock prior to reaction, considered to outweigh the energy released by dry torrefaction if the feedstock has more than 30% water[8]. Wet torrefaction, also known as hydrothermal pretreatment or hydrothermal carbonization (HTC), is usually performed at temperatures higher than 180 °C, at pressures high enough to ensure liquid water, and under an inert atmosphere. Reaction time has been reported to be 5 minutes to several hours, although most of the reaction seems to occur within the first 20 minutes[9-12]. Additives, such as acids or bases, can affect the products formed. At an initial pH greater than 7, no solid product is formed[13,14]. Adding acid to the reactants can increase the HHV of the solid product formed, as well as reduce the mass yield[15]. However, to keep the process straightforward, only undried biomass and water need be used. As both reactant and solvent, water shows different physical and chemical properties depending on the operating conditions[14]. At temperatures between 227 and 327 °C, water may act as both a base and an acid because its ionic product is maximized.

30

In addition, water’s dielectric constant is decreased at these temperatures so that it acts more like a non-polar solvent[16]. During hydrothermal pretreatment, hemicelluloses and cellulose are hydrolyzed to oligomers and monomers[17-20], while lignin is mostly unaffected. The solid product, also known as biochar, has reduced equilibrium moisture content, so it is less likely to rot in transport[21]. The pretreated solid is quite friable and might be made into pellets which can be fed to a gasifier or coal power plant easily. Mass and energy balances of the wet torrefaction process previously have been accomplished[7]. The objective of this study is to characterize the solid products produced by wet torrefaction, looking at different types of biomass, and to investigate the effect of process conditions on the quality of the solid product. 2.2 Experimental Section 2.2.1. Materials Biomass feedstocks were acquired for testing, including loblolly pine (Alabama, USA), pelletized corn stover (Pennsylvania, USA), Tahoe mix (a mixture of Jeffrey pine and white fir from the Tahoe forest), switch grass (Nevada, USA) and rice hulls (California, USA). 2.2.2 Wet torrefaction Wet torrefaction of lignocellulosic biomass materials was performed in a 100 mL Parr bench-top reactor (Moline, IL) at temperatures ranging from 200°C to 280 °C. The temperature of the reactor was controlled using a Single Display Proportional-IntegralDerivative (PID) Controller (Winona, MN).The reactor pressure was autogenic and indicated by a pressure gauge to range from 1.4 MPa to 6.9 MPa. For each run, a mixture

31

of biomass feedstock and water was loaded into the reaction vessel. The water to biomass mass ratio was either 5 or 10 to1. Prior to the reaction, the biomass was stirred manually to ensure complete wetting. Nitrogen was passed through the reactor for 10 min to purge oxygen out of the reactor. The reactor was heated up in 15-30 min and maintained at the desired temperature for a chosen period of time. Then, the reactor was cooled to room temperature rapidly by immersion in an ice bath. The gas sample was released without further evaluation. The liquid sample and solid sample were separated by filtering. All experiments were performed twice. 2.2.3 Ultimate analysis Solid samples were dried in an oven at 105 ˚C for 24 hours. Ultimate analysis of both raw biomass feedstock and pretreated biomass was performed with a FlashEA 1112 (Pittsburgh, PA) elemental analyzer for full determination of C, H, N, S, and O using the standard practice for ultimate analysis of coal and coke ASTM D3176 - 89(2002). 2.2.4 Heat of combustion Heat of combustion of solid samples was measured in a Parr 1241 adiabatic oxygen bomb calorimeter (Moline, IL). Solid samples (0.5 – 1.0 g) were dried at 105 °C for 24 h prior to analysis. 2.2.5 Fiber analysis The van Soest method of NDF-ADF-ADL (neutral detergent fiber, acid detergent fiber, acid detergent liquid) dissolution was used to determine the percentage of hemicellulose, cellulose and lignin in solid samples[23]. The contents of hemicellulose,

32

cellulose, and lignin were calculated from the difference of NDF, ADF, ADL and ash. Sample mass that is not assigned to one of those fractions consists of aqueous solubles polysaccharides. Samples were dried at 105 °C for 24 h prior to fiber analysis. 2.2.6 Scanning Electron Microscopy A FE-SEM Hitachi Scanning Electron Microscope (SEM) model S-4700 was used to visualize pretreated loblolly pine. A size of 0.5 mm loblolly pine samples of raw and pretreated at 2000C, 2300C, and 2600C was used for SEM images. The samples were maintained on special studs and platinum coated with polaran coater tar 5000, under an argon atmosphere for a coating thickness of approximately 1000 Å. The samples were dried in a 1050C oven for 24 h prior to analysis. A voltage of 15-20 kV with a magnification of 250-1000 times was used for the images. 2.3 Results and discussion 2.3.1 Effects of reaction temperature, holding time, water to biomass ratio, and particle size A factorial experimental design was used (see Table 1) to study the effects of four variables at two levels, namely, reaction temperature (T), holding (reaction) time (H), water/biomass ratio (R), and biomass particle size (S)[24]. Loblolly pine was selected as the lignocellulosic biomass feedstock. Mass yield, energy densification ratio, and energy yield are the three important measures in this study. Mass yield is defined as the mass ratio of dried pretreated solid to dried biomass times 100%. The energy densification ratio is the ratio of the higher heating value (HHV) of the

33

pretreated dried solid fuel product to that of the original dried biomass. The energy yield is defined as the mass yield times the energy densification ratio, and shows how the total fuel value of the solid product relates to the fuel value of the original biomass. With the change of reaction temperature the effect on mass yield, energy densification ratio, and energy yield is significant. For the 5 min reaction time with 0.16 mm sized samples with water biomass ratio 5, the mass yield at 2000C is 86.13%. But applying the same conditions in higher temperature at 2600C, mass yield decreases to 56.99%. By doubling the water biomass ratio, the results do not vary significantly. The same characteristics can be observed with reaction time and particle size too.

T

H

R

S (mm)

Mass Yield (%)

Energy Dens. Ratio

Energy Yield

(0C)

(min)

(g pine/g H2O)

200

5

5

0.16

86.13

1.109

95.52

200

5

10

0.24

87.16

1.073

93.52

200

20

5

0.24

85.71

1.148

98.40

200

20

10

0.16

78.98

1.115

88.06

260

5

5

0.24

56.99

1.435

81.78

260

5

10

0.16

53.64

1.236

66.30

260

20

5

0.16

49.99

1.421

71.04

260

20

10

0.24

41.97

1.475

61.91

Table 2.1: Effect of temperature, holding time, water biomass ratio, and particle size on mass yield, energy densification, and energy yield.

34

Results are shown in Table 1, and a statistical analysis (ANOVA) of the results shows that at 95% confidence, reaction temperature is the most significant variable that affects the mass yield and energy densification ratio. Other variables (holding time, water to biomass ratio and biomass feedstock size), do not appear to have a significant effect on the product distribution or quality of the pretreated solid samples. Rogalinski and co workers similarly found particle size and biomass concentration to have no effect on mass yield[25]. The statistical analysis also shows that the interactions among the four variables are not significant. Therefore, the manipulation of these variables can be done independently within the variable range of the experiments. 2.3.2 Change in biomass and temperature effects on mass yield, energy densification ratio, and energy yield

Energy Yield(%)

Mass Yield(%) T(0C)

T(0C)

Figure 2.1: Effect of temperature on mass yield (a), energy yield (b) for loblolly pine, pelletized corn stover, Tahoe mix (a mixture of Jeffrey pine and white fir from the Tahoe forest), switch grass, and rice hulls. Due to the significant effect of reaction temperature, wet torrefaction of various types of biomass was performed at temperatures of 200°C, 230°C, and 260°C, using a

35

water to biomass ratio of 5 to 1 and a reaction time of 5 min. Fig. 2.1 shows the mass yield (a), and energy yield (b) results for rice hulls, corn stover, Tahoe pine (a mix of Jeffrey pine and white fir), switch grass, and loblolly pine. Figure 2.2 shows the energy densification ratio results for rice hulls, corn stover, Tahoe pine (a mix of Jeffrey pine and white fir), switch grass, and loblolly pine. To explain the different results from the diverse biomass, looking at their constituent parts is necessary. Biomass consists of 5 main components: hemicelluloses (pentosans), cellulose, lignin, aqueous solubles, and ash. Table 2a shows fiber analysis data for each of these biomass. The relationship between biomass component percentages and mass yield is difficult to discern at the lower torrefaction temperatures, since mass yield varied little.

Energy Densification Ratio T(0C)

Figure 2.2: Effect of temperature on energy densification ratio for loblolly pine, pelletized corn stover, Tahoe mix (a mixture of Jeffrey pine and white fir from the Tahoe forest), switch grass, and rice hulls.

36

However, with 260°C pretreatment, those biomass with higher percentages of hemicellulose show lower mass yield. For example, switch grass and corn stover have more hemicellulose compared to the other biomass tested and exhibited a lower mass yield. This indicates that hemicellulose is more easily reacted in the wet torrefaction process. A lower reaction temperature would be expected to remove more hemicellulose compared to cellulose or lignin. To confirm the finding that hemicellulose is more easily reacted, fiber analysis was completed of dried raw loblolly pine and dried loblolly pine pretreated at 200 and 260°C. Table 2b summarizes the distribution of hemicellulose, cellulose, lignin, and aqueous solubles plus ash of the loblolly pine feedstock and pretreated solids.

Biomass

Hemicelluloses

cellulose

Lignin

water extractives

Ash

Switch grass

31.2

35.5

5.5

21.8

7.4

Corn stover

26.3

29.7

11.3

26.3

6.5

Tahoe mix

18.2

50.7

23.8

6.1

1.3

Rice hulls

14.9

39.8

11.3

12.9

21.1

Loblolly

11.5

55.4

30.0

2.7

0.4

Table 2.2a. Fiber analysis of raw switch grass, corn stover, Jeffery pine, rice hulls, and loblolly pine. The experimental conditions of the treatment at 200 ˚C were a water to biomass ratio of 5 to 1, a reaction time of 5 min, and a particle size of 0.16mm, and those of the 260 ˚C run were the same except that particle size was 0.24 mm. The mass yield for 200

37

˚C was 86.13% and for 260 ˚C was 56.99%, indicating that the higher temperature removed more of the biomass material. The results in Table 2b show that at 200˚C and 260 ˚C, most of the loblolly pine’s hemicellulose is extracted, and likely hydrolyzed to monosaccharides (xylose, arabinose, galactose and mannose). In addition, accounting for the mass yield of solid product, relative to the component in raw biomass, only 19% of cellulose is reacted with the 200 ˚C reaction temperature, compared to 62% of cellulose with 260˚C. The reacted cellulose would be hydrolyzed to glucose or oligomers. A preliminary analysis of the liquid products by HPLC, not shown here, shows the presence of various sugars (xylose, glucose, arabinose, galactose and mannose), consistent with other investigation[26,27].

L. pine condition

Hemicelluloses (%)

Cellulose (%)

Lignin (%)

Aqueous soluble(%)

Ash %

raw

11.5

55.4

30.0

2.7

0.4

200°C pretreated

1.1

52.1

31.0

15.3

0.5

260°C pretreated

0.7

36.9

43.5

18.1

0.8

Table 2.2b. Fiber analysis of loblolly pine pretreated at 200˚C and 260˚C. From Table 2.2b, it is shown that only 11% of the lignin was extracted at the 200 ˚C reaction temperature, and only 17% of the lignin at 260 ˚C, indicating that lignin is relatively inert in wet torrefaction. Hemicellulose and cellulose have a lower HHV compared to lignin[28]. When they are removed, a higher proportion of lignin must remain (Table 2.2b), causing a higher HHV. This fact may explain why loblolly pine,

38

which has the highest cellulose content of 55.4%, has an energy densification of 1.43, the highest of the biomasses investigated (Fig. 2.2 and Table 2.2a). After conversion to glucose, glucose can isomerize in the presence of acid, which is produced in the reaction scheme, to fructose, which dehydrates to 5-hydroxyfurfural (5-HMF).[28] 5-HMF may precipitate into the pores of the solid product and then be accounted for as aqueous solubles[16]. 5-HMF has a greater HHV than cellulose or glucose[30], and so its increased presence may partially account for the greater energy densification found in those biomass with a higher aqueous solubles percentage. For example, Tahoe mix, a mixture of pine and fir, shows a relatively low energy densification of 1.12 at a 260˚C pretreatment temperature, despite its relatively high lignin content. This could relate to its low aqueous solubles content of 6%. Switch grass and corn stover have more hemicellulose compared to the other biomass tested. Hemicellulose hydrogen bonds to cellulose, reinforcing it to prevent it from decomposing[31]. When the hemicellulose is a larger amount of the biomass structure, and is removed, pretreatment conditions likely are capable of removing the now unprotected cellulose[25]. Again, a higher proportion of lignin must remain. This may account for the increased HHV and thus the increased energy densification that switch grass and corn stover display, as shown in Fig. 2.1. and fig.2.2 In summary, pretreatment appears to increase energy densification more in biomass with a higher proportion of lignin and possibly aqueous solubles, while it decreases mass yield more in biomass with a greater proportion of hemicellulose.

39

2.3.3 Increased aqueous solubles and friability Increased pore area occurs due to the removal of cellulose and hemicellulose in the solid product. Since more surface area is available for precipitation, more sugars and other chemicals adhere to the accessible pores. For this reason, aqueous solubles increase with higher pretreatment temperature, as shown in Table 2.2b. The mass fraction of aqueous solubles compounds increases with extent of reaction, from 2.7% for raw biomass, to 15.3% for biomass pretreated at 200 ˚C, to 18.1% for biomass pretreated at 260 ˚C. When the solid is dried, it leaves behind a precipitate on the extensive surface area. Thus, the mass of these precipitates is represented as a solid, but is analyzed as aqueous solubles by fiber analysis. With vigorous washing, these compounds might be more correctly characterized as dissolved chemicals, including sugars. The empty spaces left in the solid product are also likely the cause of the higher friability of the pretreated solid product. 2.3.4 Scanning electron microscope images Fig. 2.3 shows SEM images of raw loblolly pine and loblolly pine pretreated at 200°C, 230°C, and 260°C using a water to biomass ratio of 5 to 1 and a reaction time of 5 min, and a typical construction for a plant cell wall. In general, cellulose can be found as a bundle in the raw lignocellulosic biomass, while hemicellulose is the spiral cover of the cellulose and lignin separates the bundles from each other like a stratum (Fig. 2.3(e)). Aqueous solubles can be found as a thin layer over the whole configuration. In Fig. 2.3(a), which is the SEM image of 250 times magnification of the raw loblolly pine, the lignocellulosic structure is not clear. It may be the evidence of the presence of aqueous

40

solubles. In the image of biochar pretreated at 2000C in Fig. 2.3(b), it is possible that the hemicellulose wrapping of the cellulose bundle is observable.

(a)

(b)

(c)

(d)

(e)

Fig. 2.3 Scanning electron microscope images of (a) raw loblolly pine and loblolly pine pretreated at (b) 200˚C, (c) 230˚C, and (d) 260˚C with the magnification of 250 times. (e) Possible arrangement of chemical components in lignocellulosic biomass.

41

This idea suggests that at 2000C all the aqueous solubles are completely decomposed and that hemicellulose begins decomposing. With a higher magnification of 500 times in Fig. 2.3(g), it appears that the hemicellulose is hydrolyzed and cellulose is beginning to decompose as well. In Fig. 2.3(c), where biochar pretreated at 2300C is displayed, there are no spiral objects observed, possibly hemicellulose is completely decomposed and cellulose is partly decomposed, too. However, the hollow boundary in the 500x magnification of the image at the same conditions, (Fig. 2.4(c)), may indicate that the cellulose is not completely decomposed.

(a)

(b)

(c)

(d)

Fig. 2.4 Scanning electron microscope images of (a) raw loblolly pine and loblolly pine pretreated at (b) 200˚C, (c) 230˚C, and (d) 260˚C with 500 times magnification.

42

Cellulose appears to be breaking off the biochar pretreated at 2600C in Fig. 2(d). With the temperature increase, the plane that separated the cellulose bundles from each other seems to show no change. That may be because the lignin behaves as an inert. In the 500x magnification of the same condition in Fig. 2.4(d), there are some cracks on the walls. Lignin starts decomposing at about 2600C. Kobayashi et al. (2009) reported that below 2000C, fibrous materials are observed, which suggests that the decomposition of woody biomasses begins above 2000C[32]. With an increase of temperature from 2003000C, the amount of fibrous materials decreases and the fibrous materials are converted to round shaped particles, which implies that cellulose content decreases as temperature increases. 2.3.5 Recycling of liquid product After wet torrefaction, the liquid product can be collected and reused to treat fresh biomass. Two such experiments were done with loblolly pine, one with two torrefactions at 200˚C, and the other with two at 260˚C. At 200˚C, the mass yield and energy densification ratio were relatively constant when the liquid product was re-used instead of water, but at 260 ˚C, there were some distinct changes. The mass yield increased from 57.0 % to 65.6 % and the energy densification ratio decreased from 1.435 to 1.305. The torrefaction chemistry is apparently inhibited by compounds produced at 260˚C which are water soluble, possibly because equilibrium reactions occur in the reaction scheme. Similar reduction in acetic acid production has been reported by Lynam and co-workers when acetic acid was added to initial reactants in wet torrefaction, indicating the likelihood of equilibrium reactions where acetic acid is involved[15].

43

2.3.6 Ultimate analysis As summarized in table 2.3, the ultimate analysis of loblolly pine, rice hull, and corn stover shows the increase of carbon content and decrease of oxygen content with pretreatment. All the pretreatment was done in 275°C. The decrease of oxygen content is more in loblolly pine than rice hull and corn stover with the pretreatment. Similarly it increases carbon content higher for loblolly pine than other two. Ayhan (2003) also confirmed that heating value increases with increasing lignin content and fixed carbon content in a specific range[33]. The van Krevelen diagram shown in Fig. 3(b) shows the H/C and O/C atomic ratios from ultimate analysis of raw biomass and biomass pretreated at 275 °C.

Biomass

C (%)

H (%)

N (%)

S (%)

O (%)

Loblolly

Raw

51.6

5.7

0.07

0.05

42.3

pine

Pretreated

68.9

5.2

0.12

0.03

25.5

Raw

38.9

4.6

0.26

0.08

35.6

Pretreated

43.6

4

0.36

0.05

24.2

Corn

Raw

43.3

5.3

0.72

0.04

40.3

stover

Pretreated

49.8

4.6

0.83

0.13

29.3

Rice hull

Table 2.3: Ultimate analysis of different biomass both raw and pretreated at 2750C. Wet torrefaction at 275°C moves the fuel value of the raw biomass from the top right of the diagram to a fuel value similar to peat for rice hulls or corn stover, or to

44

lignite, a low rank coal, for loblolly pine. This coalification is due to the biomass’s increased carbon content and decreased oxygen content. This is consistent with the increase in fixed carbon found in the proximate analysis and with the enhanced HHV of the solid product.

Figure 2.5: Van Krevelen diagram of loblolly pine, rice hulls, and corn stover showing effect of 275˚C wet torrefaction pretreatment. 2.4 Conclusion Wet torrefaction is a promising method for producing a homogeneous solid fuel for subsequent thermochemical conversion from diverse lignocellulosic biomass feedstocks. The energy density of solid products can be increased significantly. Reaction temperature appears to be the only process variable significantly affecting the product distribution and energy densification. The solid product has increased fixed carbon, decreased volatiles,

45

and on a molecular scale, increased carbon and significantly decreased oxygen content. Fiber analysis confirms that hemicellulose hydrolyzes at low temperatures and that as hemicellulose and cellulose are removed inert lignin remains and increases the biochar’s higher heating value. Ultimate analysis of HTC shows the pretreated solid has the characteristics of low grade coal.

46

2.5 References [1] Perlack R., Wright L., Turhollow A., Graham, R., Stokes, B., Erbach, D., 2005. Biomass as a feedstock for a bioenergy and bioproducts industry: the technical feasibility of a billion-ton annual supply. US Department of Energy, DC, USA. [2] Zuwala, J., Sciazko, M., 2010. Full-scale co-firing trial tests of sawdust and bio-waste in pulverized coal-fired 230 t/h steam boiler. Biomass Bioenergy 34, 1165-1174. [3] Bridgwater A.V., Double J.M., 1991. Production costs of liquid fuels from biomass. Fuel 70, 1209-1224. [4] Prins, M.J., Ptasinski, K.J., Janssen J.J.G.F., 2006a. More efficient biomass gasification via torrefaction. Energy 31, 3458-3470. [5] Prins, M.J., Ptasinski, K.J., Janssen J.J.G.F., 2006b. Torrefaction of wood – part 1. Weight loss kinetics, part 2. analysis of products. J. Anal. Appl. Pyrolysis 77, 28-40. [6] Sadaka, S., Negi, S., 2009. Improvements of biomass physical and thermochemical characteristics via torrefaction process. Environ. Prog. Sustain. Energy 28, 427-434. [7] Yan, W., Hastings J.T., Acharjee, T.C., Coronella, C.J., Vasquez, V.R., 2010. Mass and energy balance of wet torrefaction of lignocellulosic biomass. Energy Fuels 24, 4738-4742. [8] Libra, J. A., Kyoung S. R., Kammann, K., Funke, A., Berge, N. B., Neubauer, Y., Titirici, M-M., Fühner,C., Bens, O., Kern,J., Emmerich, K-H., 2011. Hydrothermal

47

carbonization of biomass residuals: a comparative review of the chemistry, processes and applications of wet and dry pyrolysis. Biofuels 2, 89-124. [9] Funke, A.; Ziegler, F. Hydrothermal Carbonization of Biomass: A Literature Survey Focusing on Its Technical Application and Prospects, 2009.17th European Biomass Conference and Exhibition, Hamburg, Germany. [10] Mosier, N., Wyman, C., Dale, B., Elander, R., Lee, Y. Y., Holtzapple, M., Ladisch, M., 2005. Optimization of pH controlled liquid hot water pretreatment of corn stover. Bioresour. Technol. 96, 673-686. [11] Lu, X., Yamauchi, K., Phaiboonsilpa, N., 2009. Two-step hydrolysis of Japanese beech as treated by semi-flow hot-compressed water. J. Wood Sci. 55, 367-375. [12] Knezevic, D., van Swaaij, W., Kersten, S., 2010. Hydrothermal conversion of biomass. II. Conversion of wood, pyrolysis oil, and glucose in hot compressed water. Ind. Eng. Chem. Res. 49, 104-112. [13] Hu, B., Yu, S.H., Wang, K., Liu L., Xu, X.W., 2008. Functional carbonaceous materials from hydrothermal carbonization of biomass: an effective chemical process. Dalton Trans. 40, 5414-5423. [14] Ando, H., Sakari, T., Kobusho, T., Shibata, M., Uemura, Y. Hatate, Y., 2000. Decomposition behavior of plant biomass in hot-compressed water. Ind. Eng. Chem. Res. 39, 3688-3693.

48

[15] Lynam, J. G., Coronella, C. J., Yan, W., Reza, M. T., and Vasquez, V. R., 2011. Acetic acid and lithium chloride effects on hydrothermal carbonization of lignocellulosic biomass. Bioresour. Technol. doi:10.1016/j.biortech.2011.02.035. [16] Yu, Y., Lou, X., Wu, H., 2007. Some recent advances in hydrolysis of biomass in hot-compressed water and its comparisons with other hydrolysis methods. Energy Fuels 22, 46-60. [17] Sun Y., Cheng J., 1992. Hydrolysis of Lignocellulosic material from ethanol production. Bioresour. Technol. 39, 107-115. [18] Petersen, M. O., Larsen, J., Thomsen, M. H., 2009. Optimization of hydrothermal pretreatment of wheat straw for production of bioethanol at low water consumption without addition of chemicals. Biomass Bioenergy 33, 834-840. [19] Funke, A., Ziegler, F., 2010. Hydrothermal carbonization of biomass: A summary and discussion of chemical mechanisms for process engineering. Biofuels Bioproducts & Biorefining-Bioref. 4, 160-177. [20] Bobleter O. Hydrothermal degradation of polymers derived from plants. Prog Polym Sci 1994; 19: 797-841. [21] Acharjee, T. C., Coronella, C. J., Vasquez, V. R., 2011. Effect of thermal pretreatment on equilibrium moisture content of lignocellulosic biomass. Bioresource Tech. 102, 4849-4854.

49

[22] Ehrman, T., 1994. Standard method for ash in biomass. NREL Laboratory Analytical Procedure LAP-005. [23] Goering HK, Van Soest PJ., 1970. Forage fiber analysis, USDA Agric. Handbook 379, 1-9. [24] Hicks, R.C., Turner V.K., 1999. Fundamental concepts in the design of experiments. Oxford University Press, New York. [25] Rogalinski, T., Ingram, T., Brunner, G., 2008. Hydrolysis of lignocellulosic biomass in water under elevated temperatures and pressures. J. Supercrit. Fluid 47, 54-63. [26] Matsumura Y, Yanachi S, Yoshida T, 2006. Glucose Decomposition Kinetics in Water At 25 MPa in the Temperature Range of 448−673 K. Ind. Eng. Chem. Res. 45, 1875-1879. [27] Sasaki M., Adschiri, T., Arai, K., 2003. Fractionation of sugarcane bagasse by hydrothermal treatment. Bioresour. Technol. 86, 301-304. [28] Dumitriu, S., 2005. Polysaccharides: structural diversity and functional versatility, second ed. Marcel Dekker 270 Madison Ave, New York, NY 10016. ISBN 0-82475480-8. [29] Huber, G. W., Iborra S., Corma A., 2006. Synthesis of transportation fuels from biomass: Chemistry, catalysts, and engineering. Chemical Reviews 106, 4044-4098.

50

[30] Verevkin, S.P., Emel’yanenko, V.N., Stepurko, E.N., Ralys, R.V., Zaitsau, D.H., 2009. Biomass-derived platform chemicals: thermodynamic studies on the conversion of 5-hydroxymethylkfurfural into bulk intermediate. Ind. Eng. Chem. Res. 48, 10087-10093. [31] Hayashi, T., Kaida, R., 2011. Functions of xyloglucan in plant cells. Molecular Plant 4, 17-24. [32] Kobayashi, N.,Okada, N., Hirakawa, A., Sato, T., Kobayashi, J., Hatano, S., Itaya, Y., Mori, S., 2009. Characteristics of Solid Residues Obtained from Hot-CompressedWater Treatment of Woody Biomass. Ind. Eng. Chem. Res. 48, 373-379. [33] Ayhan D., 2003. Relationships between heating value and lignin, fixed carbon and volatile material contents of shells from biomass products. Energy Source 25, 629-635.

51

Chapter 3

Reaction Kinetics in Hydrothermal Carbonization of Lignocellulosic Biomass

Hydrothermal carbonization is a pretreatment process to convert diverse feedstocks to homogeneous energy-dense solid fuel. The resulting solid fuel has favorable properties for thermochemical conversion, including high carbon content, reduced volatiles content, and is relatively friable and hydrophobic. For industrial scale implementation, knowledge of reaction kinetics is necessary for process design and reactor optimization. In this study, we report the results of experimental measurements of kinetics of hydrothermal carbonization of lignocellulosic biomass. A novel two-chamber reactor was utilized to perform hydrothermal carbonization isothermally for specified reaction times. The hydrothermal carbonization of loblolly pine was performed at various reaction times at 200, 230, and 260°C. The mass of the solid product decreases rapidly, and the fuel value increases rapidly, both during the first 2 minutes of reaction. A simple reaction mechanism is proposed and tested against the experimental results. Hemicelluloses and cellulose both degrade in parallel first-order reactions, with activation energies of 28.6 kJ/mol and 77.4 kJ/mol, respectively. In conclusion, hydrothermal carbonization can be performed much more quickly than previously thought.

52

3.1 Introduction In 2008, US domestic daily oil production was 8.5 million barrels while the daily demand was 19.5 million barrels[1]. Better awareness of finite supplies of fossil fuel and increasing awareness of climate change motivate society to identify new sustainable resources for renewable fuel. For instance, the Energy Independence and Security Act (EISA) of 2007 mandates that at least 36 billion gallons of renewable fuel be produced and used in U.S. by 2022[2-4]. The feedstock for first-generation biofuels (fuel ethanol from starch and biodiesel from oil/fat) is inadequate due to direct competition with human and animal food. Because of the abundant renewable supply of lignocellulosic biomass, much research and development has been conducted to find economic means to utilize lignocellulosic biomass as feedstock for fuel, and energy. Although lignocellulosic biomass is cheap, challenges, including diverse feedstocks, widely dispersed production, low fuel value, and seasonal availability, make the feedstock handling and transportation expensive[5]. Moreover, the chemical properties of lignocellulosic biomass make it unfavorable in traditional thermochemical applications. To overcome these challenges, there is a need for a process to homogenize the feedstocks, while simultaneously producing a stable, energy-dense, solid fuel. In the process hydrothermal carbonization, also known as "hydrothermal pretreatment" and "wet torrefaction", biomass is treated in hot compressed water, resulting in three products: gases, aqueous chemicals, and solid product. The temperature is in the range of 200–300oC, and the reaction pressures are always greater than the water vapor pressure to ensure liquid water. The solid product contains about 55–90% of the

53

mass and 80–95% of the fuel value of the original feedstock, depending primarily on reaction temperature[6,10]. The gas product is about 10-20% by mass of the feedstock, and the aqueous chemicals, primarily sugars, furfurals, and organic acids, make up the balance[6]. The dried solid product produced by hydrothermal carbonization has a higher energy density than the starting biomass feedstock. Moreover, the solid product has favorable properties for thermochemical conversion[6]. For instance, the solid product is friable, so its size can be easily reduced into the desirable ranges with less energy consumption. The solid product has much less equilibrium moisture content in an environment of constant humidity than does raw biomass[13], and it is likely to be stable for long-term storage. Furthermore, the solid product has a greater ratio of carbon to oxygen than the original biomass feedstock, meaning that the solid fuel product becomes chemically similar to coal, of great importance for applications of either gasification or co-firing with coal[6]. Hydrothermal carbonization has the great potential to overcome the technical and economical barriers utilizing lignocellulosic biomass as the feedstock for fuels, chemicals and energy. The kinetics of hydrothermal carbonization provide an important solid basis for designing a continuous hydrothermal carbonization process as well as performing the economic evaluation. Reaction kinetics of decomposition of cellulose and biomass had been investigated using small scale batch reactor or liquid-phase thermogravimetry[7-8]. However, no method has been published to ameliorate mass loss during the heating-up procedure. Hence, from the point view of isothermal reaction, there is no literature

54

addressing the kinetic analysis of hydrothermal carbonization. Therefore, more research is required to obtain the understanding of the chemical and physical processes occurring in hydrothermal carbonization, especially reaction mechanism and reaction kinetics. This study focuses on the determination of the mass loss kinetics of hydrothermal carbonization of wood by experiments in a specially-designed two-chamber reactor. The hydrothermal carbonization of loblolly pine (typical softwood lignocellulosic biomass) was performed at various reaction times for temperatures ranging from 200 to 260oC. The changes of mass yield and fuel density were obtained with increasing reaction times. Finally, the simple reaction mechanism, two parallel first-order reactions, was proposed and validated, and kinetic parameters were also obtained, which is very useful in making recommendations for industrial hydrothermal carbonization process conditions, such as temperature and residence time. 3.2 Experimental Section 3.2.1 Biomass Loblolly pine (Alabama, USA) was used as a typical lignocellulosic biomass. On a mass basis, it consists of 11.9 % hemicelluloses, 54.0 % cellulose, 25.0 % lignin, 8.7 % water-extractives and 0.4 % ash.[6] Pine sample was milled to the particle size range (0.3-0.7 mm), and dried at 105 °C for 24 h prior to the experiment. 3.2.2 Hydrothermal Carbonization Figure 3.1 shows a schematic of the reaction system, including a two-chamber reactor, a radiant heater, a temperature indicator, and a Proportional-Integral-Derivative

55

(PID) temperature controller. A two-chamber reactor is designed and built particularly for this kinetic measurement. 316 SS stainless steel was the construction material for the reactor. Both the chambers were selected 0.5 inch nominal diameter of 316 SS.

2

5 3 4 6 1

7

T2 T1

(a)

8

(b)

Figure 3.1 Two chamber reactor (a), schematic of the two-chamber reactor system (b), the components: 1. Bottom chamber; 2. Top chamber; 3. Ball valve; 4. Pressure relief valve; 5. Water-cooling coil; 6. Radiant heater; 7. Temperature indicator; 8. PID temperature controller.

56

The bottom chamber (volume: 20 mL) and the top chamber (volume: 10 mL) are connected with a Swagelok ball valve (Sunnyvale, CA), which can handle high temperature (up to 454oC) and high pressure (up to 1000 psi). Ceramic radiant heater from Omega (Stamford, CT) is applied to heat the bottom chamber of the reactor, where hydrothermal carbonization reaction actually occurs. Since there is a constant temperature gradient (90oC) between the chamber-wall temperature and the chambercenter temperature, a PID temperature controller from Omega (Stamford, CT) is utilized to achieve an accurate control of chamber-wall temperature.

Figure 3.2: Making of cigarette-shape capsule sample holder. Biomass sample is stored in the top chamber while the bottom chamber heats to the desired temperature. Both the water-cooling coil placed on the top chamber and the ball valve serves to maintain the biomass sample at the room temperature during the heating-up period. From the point view of the instrumental safety, a pressure relieve

57

valve is also installed in the two-chamber reactor, with release pressure set at 2.5 MPa for 200 and 2300C pretreatments and 6 MPa for 2600C pretreaments.

Figure 3.3: Method for placing biomass in hot compressed water for rapid sample heating. (Left to right) (a) water is heated while sample in top chamber, (b) open the ball valve allowing sample to drop into the bottom chamber, (c) sample is immersed in the hot water and ball valve is closed. The experimental procedure of hydrothermal carbonization is as follows: 15 mL of de-ionized water is firstly loaded into the bottom chamber, and the ball valve is 95 % closed. The biomass sample (0.2 g) is wrapped into the close-ended cigarette-shape capsule, which is made by the stainless steel mesh (320 mesh) from TWP (Berkeley, CA). The stainless steel capsule (fig.3.2) ensures all biomass sample stays intact. The biomass

58

capsule is placed into the top chamber. Nitrogen is passed though the reactor for 10 min to remove oxygen out of the reactor. Then the ball valve is fully closed. The bottom chamber starts heating (fig.3.3.a), which takes approximately 20-30 min, depending on the reaction conditions. Once the chamber-center temperature reaches the desire temperature (25oC higher than hydrothermal carbonization temperature), the ball valve was fully opened (fig.3.3.b), then closed in 3 seconds in order to drop the biomass capsule into the bottom chamber. Then hydrothermal carbonization reaction immediately starts isothermally at the desired temperature (fig.3.3.c). After a certain period of time (e.g. 15 s, 30 s, 45 s, 1 min, 2 min, 3 min, 4 min, 5 min), the reactor was quickly removed from the radiant heater and immersed into a ice-water bath to quench the reaction. Once the reactor reaches the room temperature, biomass capsule is removed out of the bottom chamber, and washed off aqueous chemicals absorbed by the pretreated biomass. The wet pretreated biomass was dried at 105oC for 24 h for further analysis. All experiments are performed at least 3 times and data are reported as the average. 3.2.3 Fuel Content Higher heating value (HHV) for biomass and pretreated biomass was measured in a Parr 1241 adiabatic oxygen bomb calorimeter (Moline, IL), which is fitted with continuous temperature recording system[9]. 3.3. Results and Discussion 3.3.1 Hydrothermal Carbonization of Loblolly Pine

59   

Hydrothermal carbonization of loblolly pine was performed in hot compressed water at the temperatures ranging from 200 to 260oC. In order to achieve accurate measurement of reaction kinetics, the reaction must be performed at the desired condition once the reaction occurs. Consequently, hydrothermal carbonization starts immediately at the isothermal condition. The experimental data are collected at eight intervals, which are 15 s, 30 s, 45 s, 1 min, 2 min, 3 min, 4 min and 5 min. The experimental results of hydrothermal carbonization of loblolly pine are summarized in Table 3.1. The mass yield, energy densification ratio, and energy yield are defined as below.

Mass yield =

Mass of dried pretreated biomass × 100% Mass of dried biomass

Energy densification ratio =

HHV of dried pretreated biomass HHV of dried biomass

Energy yield = Mass yield × Energy densificat ion ratio

The results (Table 3.1) have proven the importance of this two-chamber reactor for the kinetic study of hydrothermal carbonization since there is the significant change of mass yield of pretreated biomass in the period of the first minute. Within the first minute, the mass yields drop to 81 %, 65 % and 56 % at 200, 230, and 260°C, respectively. With further reaction time, the reaction rates become less dramatic with the increasing reaction temperature. The trend of mass loss can be more predictable in the figure 3.4. At 5 min, the mass yields reach 64 %, 58 % and 54 % at 200°C, 230°C and 260°C. This is explained by the different reactive characteristics of main components of lignocellulosic biomass.

60    Temperature (oC)

Time (s)

Mass yield

HHV

(%)

15

Energy yield

(MJ·kg-1)

Energy densification ratio

90.13±1.8

19.45±0.28

1.01

91.21

30

85.02±1.7

19.52±0.70

1.02

86.37

45

82.45±1.7

19.72±0.56

1.03

84.59

60

81.44±2.2

19.88±1.45

1.03

84.25

120

76.8±2.1

20.41±1.43

1.06

81.56

180

74.45±1.8

20.9±0.24

1.09

81.03

240

69.59±0.9

21.63±0.59

1.13

78.33

300

63.88±0.1

21.91±0.90

1.14

72.84

15

85.47±2.2

19.76±0.07

1.03

87.90

30

76.5±0.9

20.47±0.73

1.07

81.51

45

70.19±0.8

21.06±0.21

1.10

76.92

60

66.48±0.9

21.98±0.50

1.14

76.04

120

63.65±0.5

22.73±0.20

1.18

75.30

180

62.7±1.5

23.17±0.41

1.21

75.59

240

59.51±1.0

23.39±0.75

1.22

72.44

300

58.04±1.1

23.44±0.78

1.22

70.79

15

85.83±1.9

21.15±0.92

1.10

94.47

30

73.79±0.9

21.97±0.20

1.14

84.34

45

63.69±1.1

24±1.02

1.25

79.54

60

55.9±1.4

26.03±1.44

1.35

74.40

120

54.94±4.2

26.39±1.15

1.37

76.77

180

54.7±4.9

26.53±1.06

1.38

76.34

240

54.6±3.4

26.55±0.38

1.38

75.56

300

54.3±4.5

26.16±1.84

1.36

74.33

(%)

200

230

260

Table 3.1. Hydrothermal carbonization of loblolly pine. The high heating value (HHV) of raw loblolly pine is 19.22 MJ·kg-1.

61   

Figure 3.4. Mass yield of hydrothermal carbonized loblolly pine at reaction temperatures. Curves only indicate the tendency of the change of mass yield. Hemicellulose is very reactive in hot compressed water at even low temperatures (e.g. 180°C). It is possible that hemicellulose has been decomposed into five-carbon sugars and six-carbon sugars at 200°C. Cellulose, another major component of lignocellulosic biomass is less reactive than hemicelluloses. However, the decomposition rate of cellulose increases significantly with increasing temperature. Lignin is the least reactive component and decomposition of lignin only occurs dramatically at high temperatures (e.g. 270°C).

62   

Figure 3.5. Fuel content of hydrothermal carbonized loblolly pine (on a dry basis) at reaction temperatures. The higher heating value (HHV) of raw loblolly pine is 19.22 MJ·kg-1. Curves only indicate the tendency of the change of fuel content. Hydrothermal carbonization is a promising energy densification process because after 5-min hydrothermal carbonization, the energy densification ratios are 1.14, 1.22, and 1.36 at the temperature of 200°C, 230°C and 260°C, respectively. Figure 3.5 shows the change of fuel content of pretreated biomass with reaction time at each temperature. At 200°C, the energy density of pretreated biomass increases slowly and linearly with increasing reaction time. While at 230°C and 260°C, there is a significant densification occurring in the period of first minute of hydrothermal carbonization. Combination of information from mass yield and fuel content indicate that the reaction rate could be very

63   

fast at the beginning of hydrothermal carbonization and the reaction may accomplish in a short period of time (e.g. 60s), especially at high temperatures like 260°C. 3.3.2 Kinetic Model for Hydrothermal Carbonization Owing to the molecular structure of lignocellulosic biomass, the complex reaction scheme in hydrothermal carbonization could involve hundreds of different reactions. However, hydrothermal carbonization is used as a pretreatment process to improve the physical and chemical properties of lignocellulosic biomass to make it as feedstock for energy production in an economic manner. Hence, the solid product after hydrothermal carbonization is of the most interest. The kinetics of weight loss in hydrothermal carbonization is of great importance for further process design and economic evaluation. It was reported that hemicelluloses and water-extractives can decompose quickly at low temperatures even at 200oC. The reaction rate of cellulose decomposition in hot compressed water is definitely slower than that of hemicelluloses. Compared with hemicellulose and cellulose, lignin reacts on much smaller scale even at high temperature. According to the different characteristics of major components in lignocellulosic biomass, a simply kinetic model, consisting of two parallel first-order reactions was proposed for hydrothermal carbonization of lignocellulosic biomass. One first-order reaction represents the decomposition of hemicelluloses, which results in two types of products: aqueous chemicals and gases. The other first-order reaction represents the decomposition of cellulose, which produces the solid product, aqueous chemicals as well as gases. The solid product is called biochar. Biochar could be part of the cellulose of high crystallinity remaining at the temperatures investigated herein[11].

64   

Initial value of Loblolly Components

Symbol

Reaction type pine(%)[6]

Aqueous S

8.7

Instantaneous

Hemicellulose

H

11.9

First Order

Cellulose

C

54.0

First Order

Lignin

L

25.0

Inert

Solubles

Table 3.2. Fiber Analysis of raw loblolly pine and predicted reaction type of different constituents in loblolly pine Aqueous chemicals include oligosaccharides, monosaccharides, acids, furfurals, etc. For gases, carbon dioxide account for over 90 % on a molar basis[10]. Two first-order reactions are shown in Eqs. (1) and (2).

H → Sp + G

(1)

C → β B c + (1 - β )(Sp + G)

(2)

where S, H, C, Sp, G, Bc represents aqueous extractives, hemicelluloses, cellulose, aqueous chemicals, gases, and biochar, respectively. The mass yield of biochar from

65   

cellulose is denoted as the parameter β for the second reaction. Owing to the first-order reaction, the differential rate equations are given for hemicelluloses and water-extractives, and celluloses by Eqs. (3) and (4). k1 and k2 are rate constants for two reactions.

d S ( t) dt d H ( t) dt d C ( t) dt

∞ (3) 

− k 1 ⋅ H ( t) (4)

− k 2 ⋅ C ( t)

The system of equations can be solved analytically. Integration of the differential equations, with the initial mass at the time equal to zero, give the expression for the functions of H(t), C(t), Bc(t), which are shown by Eqs. (5), (6), and (7). Since lignin is considered as inert component, function of L(t) is expressed as constant (see Eq. 8).

H(t ) = H 0 e − k1t

(5)

C(t ) = C0e − k2t

(6)

Bc (t ) = βC0 (1 − e − k2t )

(7)

L(t ) = L0

(8)

Where H0, C0, L0 represents initial mass of hemicelluloses and water-extractives, cellulose, and lignin, respectively. If M(t) represents the mass of solid biomass at time t in hydrothermal carbonization, M(t) can be written as:

66   

M(t ) = H(t ) + C(t ) + B c (t ) + L(t )

(9)

To express the mass yield of biomass Y(t), Eq. (9) can be rewritten as:

Y(t ) =

M(t ) = YH0 e − k1t + YC0 e − k 2t + β YC0 (1 − e − k 2t ) + YL0 M0

(10)

Where YH0, YC0, and YL0 represents initial mass content of hemicelluloses and waterextractives, cellulose, and lignin in lignocellulosic biomass feedstock, respectively. 3.3.3 Parameter Estimation of Mass Yield Curves for Hydrothermal Carbonization For loblolly pine, the initial mass content of hemicelluloses and water-extractives, cellulose, and lignin are 20.6 %, 54.0 %, and 25.0 %, respectively. Figure 3.3 indicates that after 5 min mass yield continues to decrease at 200oC and 230oC, while remaining constant at 260oC. To determine the mass yield of biomass, it is reasonable to predict it by Eq. (11) where Y(t→∞) is the mass yield of biomass after 5 min of hydrothermal carbonization at the temperature of 260oC. The mass yield of biochar β is equal to 0.537 for all temperatures. So, we can assume that β is temperature-independent in the range investigated in this study. To validate this assumption, hydrothermal carbonization is performed for much longer period of time, such as 10 min, 15 min and 30 min (see Figure 3.6). The results show that the mass yields of pretreated biomass at three temperatures reach the same values 54.4 % after 30 min of hydrothermal carbonization.

1 - H 0 - (1 - β )C 0 = Y(t → ∞)

(11)

67   

Figure 3.6. Model prediction of hydrothermal carbonization of loblolly pine for long reaction time sat reaction temperatures. The mass yield is calculated in equation 10, in terms of 2 unknown rate constants k1 and k2. The best fit for rate constants can be obtained by minimizing the objective function F (see Eq. 12). 8

F(k 1 , k 2 ) = ∑ (Yiexperimental −Yimodel ) 2

(12)

i =1

model Where, F represents the experimental mass yield. Yi represents the calculated mass

yield. i represents each of 8 reaction times for any given temperature. The figure 3.7

68

shows the kinetic model prediction of hydrothermal carbonization of loblolly pine, from equation 10, using rate constants given in Table 3.3.

Figure 3.7. Kinetic modeling of hydrothermal carbonization of loblolly pine at reaction temperatures. 3.3.4 Kinetic Parameters of Hydrothermal Carbonization Table 3.2 represents the kinetic parameters of hydrothermal carbonization of loblolly pine. The values of lnk1 and lnk2 for three temperatures are plotted versus inverse temperature (see Figure 3.8). It is easy to find the relationship is linear with different slopes. Activation energies and pre-exponential factors are obtained from these slopes by the Arrhenius Equation (see Table 3.2). Pre-exponential factors are 58.58 s-1 and 824.06×103 s-1 for the first and second reactions. The activation energy is 28.56 kJ·mol-1

69

for the first reaction, and 77.42 kJ·mol-1 for the second reaction. It means the activation energy for the second reaction is much higher than that for the first reaction. These findings agree with the literature[6,10], because the first reaction involves the decomposition of hemicelluloses, which is much faster than decomposition of cellulose.

Figure 3.8. Determination of kinetic parameters of hydrothermal carbonization of loblolly pine. Prins and co-workers reported the kinetic study of torrefaction of lignocellulosic biomass using two consecutive first-order reactions. The activation energy are 75.98 kJ·mol-1 and 151.71 kJ·mol-1, respectively[12]. Compared with torrefaction,

70

hydrothermal carbonization not only has much lower activation energy but also produced much more favorable solid product for further thermochemical conversion.

T

k1

E1

k01

k2

E2

k02

(oC)

(s-1)

(kJ·mol-1)

(s-1)

(s-1)

(kJ·mol-1)

(s-1)

200

0.04

230

0.07

77.42

824.06×103

260

0.09

0.0022 28.56

58.58

0.0085 0.0200

Table 3.3. Kinetic parameters of hydrothermal carbonization of loblolly pine. k1 and k2 are rate constants. E1 and E2 are activation energy. k01 and k02 are pre-exponential factors. 3.4 Conclusions Kinetics measurement and modeling indicate important characteristics of hydrothermal carbonization of lignocellulosic biomass. Mass yields of pretreated biomass drop dramatically from 100 % to 56-81 % within the first minute and continue decreasing with reaction time. After 5 minutes hydrothermal carbonization, mass yields reach 5464 % depending on the reaction temperature. The significant changes in fuel density also occur as well. After 1 minute, higher heating value (HHV) reaches 19.9, 22.0 and 26.0 MJ·kg-1 with increasing temperature, while 22.0, 23.5, 26.2 MJ·kg-1 after 5 minutes accordingly.

71

At temperatures ranging from 200oC to 260oC, hydrothermal carbonization of lignocellulosic biomass can be well presented by a simple reaction mechanism, two parallel first-order reactions. Both reactions represent the decomposition of hemicelluloses and aqueous extractives, and decomposition of cellulose, respectively. Activation energy for two first-order reactions are 28.56 kJ·mol-1 and 77.42 kJ·mol-1, which are much lower than that of wet torrefaction. Thus, the kinetic results confirmed that hydrothermal carbonization can be achieved in a very short period of time, which results in the reduction of reactor volume. Consequently, the capital and operating costs of hydrothermal carbonization can be significantly reduced to prepare diverse lignocellulosic biomass feedstock to solid fuel with favorable physical and chemical properties in thermochemical conversion.

72

3.5 References [1] International Energy Statistics, EIA, US Energy Information Administration, US Department of Energy. [2] Biomass multiyear program. Office of the Biomass Program, Energy Efficiency and Renewable Energy, U.S. Department of Energy, Washington, D.C. 2008. [3] Huber, G.W.; Dale B.E. Sci. Am. 2009, July, 52. [4] Biofuels create green jobs: Growing transportation fuels and the nation’s economy. Office of the Biomass Program, Energy Efficiency and Renewable Energy, U.S. Department of Energy, Washington, D.C., 2008. [5] Tester, J.W. Sustainable Energy, Cambridge, Massachusetts: MIT press, 2005. [6] Yan, W.; Acharjee, T.C.; Coronella, C.J.; Vasquez, V.R. Environ. Prog. Sustainable Energy 2009, 28, 435. [7] Kamio, E.; Takahashi, S.; Noda, H.; Fukuhara, C.; Okamura, T. Ind. Eng. Chem. Res. 2006, 45, 4944. [8] Mochidzuki, K.; Sakoda, A.; Suzuki, M. Adv. Environ. Res. 2003, 7, 421. [9] Operating instructions for the 1241 oxygen bomb calorimeter. Parr Instruments, Moline, IL, 203M, p1. [10] Yan, W. Hastings, J.T., Acharjee, T.C.; Coronella, C.J.; Vasquez, V.R. Energy Fuels 2010, 24, 4738-4742.

73

[11] Minowa, T.; Zhen, F.; Ogi, T. Journal of Supercr Fluids 1998, 13, 253. [12] Prins, M.J.; Ptasinski, K.J.; Janssen, J.J.G.F. Journal Anal. Appl. Pyrolysis 2006, 77, 28. [13] Acharjee, T.C., Coronella, C.J., Vasquez, V.R. Bioresource Technology, submitted August 21, 2010.

74

Chapter 4

Pellets from Pretreated Biomass

For lignocellulosic biomass, costs of biomass supply, logistics, and diversity of feedstock characteristics inhibit commercialization for advanced fuel and power production. Hydrothermal carbonization (HTC, or wet torrefaction) is a pretreatment process for making a homogenized, carbon rich, and energy dense solid fuel, called biochar, from underutilized lignocellulosic biomass. Compared to raw biomass, biochar is both more hydrophobic for better storage and more friable for better processing. In this pretreatment method, the biomass is treated with hot compressed water in an inert environment in the temperature range of 200-260°C. The range of mass yield of biochar is 60-90% and its energy yield is 75-95%, indicating that the process causes an energy densification of 6-25%. Higher reaction temperature decreases mass yield and increases energy densification. Kinetic data show that hemicelluloses and water extractives react very quickly compared to cellulose and that lignin is relatively inert, resulting in a biochar with increased lignin content. Lignin exhibits a glass transition temperature of about 1371650C, and it can act as a binder for pelletization of pretreated lignocellulosic biomass. A hydraulic press with a controllable heated die was used for pelletization. The materials used to make pellets were raw loblolly pine and pine pretreated at 200°C, 230°C, and 260°C. The pine was pressed into a 13 mm die to make cylindrical pellets 5-8 mm long.

75

The pressure applied was 20 MPa, while a temperature of 1400C was maintained for 30 seconds to allow the lignin to melt and bind each pellet. Pellets made from HTC biochar exhibit favorable properties resulting from high levels of lignin, including increased hydrophobicity, abrasion resistance, energy density, and mass density. But elastic modulus and the ultimate breaking strength by compression is lower for the pellets made from HTC biochar. Pellets produced from biomass pretreated at 260°C have a volumetric fuel value 70% greater than pellets produced from untreated pine.

76

4.1 Introduction Transportation and handling of lignocellulosic biomass, such as wood, rice hulls, straw and switch grass, are often challenging as they have low bulk density, from the range of 60-80 kg/m3 for agricultural straws and grasses and 200-600 kg/m3 for woody biomass[1-3]. Thermal or chemical pretreatment processes can produce a mass and energy dense product[4]. The storage of biomass can be facilitated by pretreatment, yet the problem is not solved as the solid product is more friable and dusty. Pelletization can also increase the mass and energy density of the biomass[3]. It reduces the transportation costs and provides for better handling and feeding. The pelletization process significantly reduces dust formation and causes the product to have a common shape and size. A particular size is necessary to feed into the boiler in the case of co-firing of biomass with coal. Lignocellulosic biomass pretreatment combined with pelletization potentially could improve storage for those seasonal crops which are harvested only a few weeks in a year[5]. There are two different thermal pretreatment technologies widely available. They are called wet torrefaction and dry torrefaction. In wet torrefaction, also known as hydrothermal carbonization, biomass is treated with hot compressed water resulting in three products: gas, aqueous solubles, and solid product or biochar. Reaction temperatures are in the range of 200-260oC, and the pressures are up to 4.6 MPa. The gas product is about 10% of the original biomass, containing mainly CO2, while the aqueous solubles are primarily sugars, acetic acid, and the other organic acids. The solid product contains about 55-90% of the mass and 80-95% of the fuel value of the original feedstock[4,6]. The other alternative, dry torrefaction or mild pyrolysis, treats biomass in

77

an inert gas environment and a temperature range of 200-300oC. The solid and gases are the two outputs from the process. The solid is about 60-80% by mass of the original feedstock, with a fuel value of 70-90%. The balance are gases. Both torrefaction processes exhibit solid products with higher energy densification that are easily friable and more hydrophobic relative to the original biomass. The reaction mechanisms are still poorly understood[4]. Although the torrefaction processes improve the biomass energy densification, it is friable and still hard to handle. The pelletization process can make torrefied biomass more uniform, dense, and easy to handle. The pelletization process depends on various properties such as temperature, moisture content, biomass type, binder, and pelletizer type, along with pressure. Mani and co-workers reported that there are three stages of densification of biomass under pressure[7].

Figure 4.1: Deformation mechanism of powder particle under pressure[33,34].

78

Fig. 4.1 shows the mechanism of binding under compression. In the first stage, particles form a close packed mass by rearranging themselves but retain their own properties. In the second stage, the particles are forced against each other by the applied pressure and plastic or elastic deformation takes place. In this stage the surface contacts becomes greater by solid bridge, van der Waal’s, electrostatic forces, and mechanical interlocking which promotes binding. In the last stage, the volume is again reduced by the applied high pressure until the maximum density is attained. The pellet can no longer change its density after that. The bonding will break up if more pressure is applied. Binders are used to ease up the second stage as it increases more surface contacts and promotes better bonding. Binders can be natural or synthetic. If there is any component in the feed that shows this characteristic, it is called a natural binder. Lignin is a one of the components of the lignocellulosic biomass, which shows binding behavior above 140oC. The amount of lignin in biomass varies from 15-22% on dry basis[8], and it has a glass transition temperature of about 140-180oC [9]. On the other hand, some synthetic components can be used externally to bind the feed materials. Starch, caustic soda and urea formaldehyde are some common binders[10,29]. A binder with a lower glass transition temperature is always preferable. However, it has to have the ability to bind the given feed material. Generally the glass transition temperature of the binder is the applied temperature for pelletization. Moisture can act as a self-binder or enhance the activity of the binder. The moisture in the biomass can increase the bonding via van der Waal’s forces[11,12]. It also fills up the pores among the particles and joins them together. Feed with comparatively low moisture (5-10%) content is preferable for

79

producing stable, hard, and durable pellets. High moisture (>15%) usually causes a lower heating value, and makes the pellet difficult to eject from the pelletizer, resulting in clogging of the die[13]. A smaller particle feed size can make high quality pellets as it increases the contact surface area. But very small size particles can jam the pelletizer[14]. It is reported that for alfalfa pellets increasing the screen size from 2.8 to 6.4 mm reduced the durability of the pellets by more than 15%[15]. L/D (length over diameter) ratio is important for the durability and strength of pellets[15]. Holding time is the time when the pelletization conditions applied on the biomass and relaxation time is the time to release the pelletization conditions. These two times affect the durability and strength of pellets. Even an increase of 10 s in holding time can increase the density of oak sawdust pellets by 5% [16]. For densification or pelletization of biomass, briquetting machines like hydraulic presses or pellet mills are very popular. In the case of hydraulic presses, energy to the piston can be transmitted from an electric motor or a manual hydraulic jack. Usually pellet presses can provide more pressure while temperature is controllable by using temperature controller with the heater[17]. The pellet quality is higher with a hydraulic press because all the main parameters can be controllable. But the throughput is lower than for other pelletizers as the cycle of the cylinder is slower. The products from hydraulic presses have higher bulk density as it can handle a higher pressure and a wider range of moisture. Abrasion index is a common test for pellet durability. The test is called the MICUM test and is specifically used for coal[18]. In this test, fixed amount of samples feed into

80

the rotating drum and rotate the drum for a certain number of revolution. After that the samples pull out from the drum and sieve it and calculate the fraction of samples under a certain size and this fraction is known as abrasion index[35]. A vibrating bed at a fixed frequency and amplitude can be another option for durability testing. The mechanical strength of the pellet can be measuring with a tensometer for compression of a pellet in the radial direction[17]. Narra and co-workers determined the strength of pellets by crushing them with ZWICK-ROELL material testing (tensile and compressive strength) machine[19]. Thermogravimetric analysis by TGA can reveal the thermal behavior and Digital Scanning Calorimeter can determine the glass transition temperature of the feed. The main goal of this work was to make pellets of wet torrefied biomass and examine the pellet quality and operating conditions needed using a hydraulic press. The solid residue of the hydrothermal carbonization (HTC) is called HTC biochar. The possibility of using lignin as a natural binder was tested. The glass transition behavior was verified and the temperature of pelletization was maintained above the glass transition temperature. Abrasion index, durability, modulus of elasticity, ultimate strength of pellets, and equilibrium moisture content (EMC) were determined and compared with the literature. 4.2 Materials and Methods 4.2.1 Biomass and Chemicals Loblolly pine (Alabama, USA) was used as a typical lignocellulosic biomass for torrefaction. The solutions and filter bags for fiber analysis were purchased from

81

ANKOM Technology Inc (Macedon, NY). A hydraulic press of 15 MT and different dies were purchased from Across International (New Providence, NJ). The heated die with controller was purchased from Across Int.( New Providence, NJ). 4.2.2 Wet Torrefaction Wet torrefaction of loblolly pine was performed in a 100 ml Parr bench-top reactor (Moline, IL) at temperatures ranging from 200-260oC[4]. The temperature of the reactor was controlled using a PID controller. The reactor pressure was not controlled but indicated by the pressure gauge and ranged from 1-5 MPa. For each run, a mixture of loblolly pine (size of 0.5 mm mesh) and water with a ratio of 1:5 w/w, was loaded into the reactor. Nitrogen of 0.5 MPa was passed through the reactor for 10 minutes to purge the oxygen. The reactor was heated up to the desired temperature and maintained at that temperature for 5 minutes. Then the reactor was cooled rapidly by immersing it in an icewater bath. The gas was released to the atmosphere. The solid output was filtered from the liquid and put into a drying oven at 105oC for 24 hours before further analysis. 4.2.3 Pelletization Technique The HTC biochar was exposed to ambient conditions for 3 weeks to equilibrate the pellet moisture content. Around 1 g of the wet torrefied biomass was fed into the 13 mm die in the hydraulic press. A band heater of 500W was used to heat the sample and a controller maintained the sample temperature at about 140oC in the die. A pressure (force per area) of 1000 MPa then was applied to the sample manually. The holding time was 30 s for this study. After the holding time the pressure was released and the heater was

82

turned off simultaneously. The pellet was pulled out from the die and left undisturbed for 2-5 min. It was then stored at room temperature before further analysis. The L/D ratio of the pellets ranged from 0.6-0.75 in this study. 4.2.4 Abrasion Index and Durability To evaluate the durability or mechanical strength of the pellets, the MICUM test, which is popular for characterization of coal, was adapted here[18]. 40 pellets were charged into a rotating drum with an inner diameter of 101.6 mm and a depth of 95 mm. Two opposite baffles of 25.4 X 88.9 mm were installed perpendicular to the cylinder wall. The rotation of the drum was selected to be 38 rotations per minute. After 3000 rotations each pellet was analyzed. After being revolved in the drum, the sample was screened using a 1.56 mm sieve. Particles smaller than 1.56 mm were then weighed. Abrasion index is the ratio of mass percentage below the 1.56 mm to the initial sample mass after 3000 rotations. The smaller the abrasion index, the better quality is the pellet. 4.2.5 Equilibrium Moisture Content The equilibrium moisture content (EMC) was measured at 300C by the static desiccators technique[31]. The solid samples were exposed to an environment with constant humidity and temperature for a long period of time, until the moisture in the solid reached an equilibrium value. The humidity in the chamber was maintained at a constant value by keeping the air in equilibrium with an aqueous solution saturated with a particular salt. LiCl and KCl salt solutions were used, with the humidities of 11.3% and 83.6% respectively. The pellets (0.9-1.2 g) were dried at 1050C for 24 h and then

83

immediately transferred into the desiccators for long enough to reach equilibrium. The weight of the pellets was measured every day and when it was same (within 1 mg) for three consecutive days, it was considered to be at equilibrium. 4.2.6 Higher Heating Value The higher heating value (HHV) of pellets was measured in a Parr 1241 adiabatic oxygen bomb calorimeter (Moline, IL) fitted with continuous temperature recording. The sample of 0.9-1.2 g was dried at 1050C for 24 h prior to analysis. 4.2.7 Digital Scanning Calorimetry Digital Scanning Calorimetry (DSC) is a useful technique to detect the glass transition behavior of the polymers. But the biomass or HTC biochar does not show the glass transition behavior precisely although it has lignin with it[31]. The amount of lignin presents in the biomass and HTC biochar may not be enough to show a significant change in heat flow with temperature in DSC. The extraction of lignin is important to detect the glass transition behavior of it[24]. The van Soest method of NDF-ADF-ADL (neutral detergent fiber, acid detergent fiber, acid detergent liquid) dissolution was used to extract lignin from the biomass and HTC biochar[20]. Samples were dried at 105 °C for 24 h prior to extraction. The final solid product of this method was lignin and ash[21]. The ash content of loblolly pine is very low,