COMPACTON OF SWITCHGRASS FOR VALUE ADDED UTILIZATION Except where reference is made to the work of others, the work described in this thesis is my own or was done in collaboration with my advisory committee. This thesis does not include proprietary or classified information. _________________________________________________ Zahra J. Colley Certificate of Approval: _______________________ _________________________ Oladiran Fasina, Co-Chair Yoon Y. Lee, Co-Chair Assistant Professor Professor Biosystems Engineering Chemical Engineering _______________________ _________________________ Christopher B. Roberts Steven Taylor Uthlaut Professor Professor Department Head Department Head Chemical Engineering Biosystems Engineering ________________________ Stephen L. McFarland Acting Dean Graduate School COMPACTON OF SWITCHGRASS FOR VALUE ADDED UTILIZATIO N Zahra J. Colley A Thesis Submitted to the Graduate Faculty of Auburn University in Partial Fulfillment of the Requirements for the Degree of Master of Science Auburn, Alabama May 11, 2006 iii COMPACTON OF SWITCHGRASS FOR VALUE ADDED UTILIZATION Zahra J. Colley Permission is granted to Auburn University to make copies of the thesis at its discretion, upon request of individuals or institutions and at their expense. The author reserves all publication rights. ____________________________ Signature of Author ____________________________ Date of Graduation iv VITA Zahra J. Colley, daughter of Barry and Dianne Colley, was born July 23, 1979, in New York City, New York. She graduated from Forrest City High School in Forrest City, Arkansas in 1997. She then attended Tuskegee University in Tuskegee, Alabama where she graduated with honors with a Bachelor of Science degree in Food Science in May, 2002. After an internship at the National Aeronautics Space Administration at Johnson Space Center in Houston, Texas, she began a graduate program at Auburn University in Chemical Engineering in August of 2002. v THESIS ABSTRACT COMPACTON OF SWITCHGRASS FOR VALUE ADDED UTILIZATION Zahra J. Colley Masters of Science, May 11, 2006 (B.S., Tuskegee University, 2002) 132 Typed Pages Directed by Oladiran Fasina and Y.Y. Lee There are increasing concerns about energy prices, availability and utilization in the world. This has led to many governmental and privately sponsored studies and research on the potential of renewable energy from biomass. Switchgrass (Panicum virgatum L.), a potential energy crop, has been evaluated and is being developed as an alternative to fossil fuels. Most lightly dense biomass such as switchgrass are easier to store, transport, and handle after they are densified/pelleted. The optimal pelletization method of switchgrass was obtained by carrying out fundamental studies on the effect of process parameters (moisture content, temperature, and die size) on pelletability, density and specific energy. Pellets were manufactured in a single die apparatus attached to a texture analyzer. Results showed that with a compaction force of 3924 N, the density of the switchgrass pellets increased with decreasing die size (4.8 mm to 7.9 mm). Density also increased with temperature from 60 to 90?C. The density of the compacts was also vi affected by the moisture content of the feed material. The density varied from 850 kg/m 3 to 1250 kg/m 3 . There was no significant effect of temperature on the specific energy used to make switchgrass pellets. Switchgrass used for investigation of moisture effect on physical properties was pelleted through a 4.8 mm diameter die. It was found that the bulk density, particle density, durability and hardness of the pellets were significantly affected by moisture content. The maximum values of bulk density and particle density were 708 kg/m 3 and 1462 kg/m 3 respectively. The force required to rupture the pellets varied from 32N at 6.32% to 22N at 17.4% moisture content. Durability of the pellets was also affected by moisture content and was the highest at 8.62% moisture content. The pellets absorbed moisture at rates that were significantly affected by relative humidity of the surrounding air (P<0.05). The equilibrium moisture content and equlibrium relative humidity (EMC-ERH) relationships for the pellets were sigmoidal in shape and were best predicted by the Chung-Pfost equation. Results from compositional analyses showed significant differences between lignin and ash for ground and pelleted switchgrass. vii ACKNOWLEDGEMENTS The author would like to sincerely thank the Chemical and Biosystems Engineering departments for this collaborative effort. The author would also like to thank Dr. Oladiran Fasina for being very patient and supportive. The author would like to thank and acknowledge Dr. Y. Y. Lee for the use of his laboratory and materials and all the fellow graduate students who assisted in the lab. Special thanks are due to family members and friends for their support during the course of this investigation. viii Style manual or journal used: Publication format Computer software used: Microsoft Word, Microsoft Excel, SAS 9.1 Statistical program ix TABLE OF CONTENTS LIST OF FIGURES ........................................................................................................... xi LIST OF TABLES............................................................................................................ xii INTRODUCTION .............................................................................................................. 1 CHAPTER 1 ?LITERATURE REVIEW ........................................................................... 3 BIOENERGY .................................................................................................................. 3 SWITCHGRASS............................................................................................................. 5 PHYSICAL PROPERTIES ............................................................................................. 7 Moisture Sorption Isotherms........................................................................................ 9 Rate of Moisture Sorption.......................................................................................... 13 COMPACTION............................................................................................................. 14 Process Variables Affecting the Compaction Process ............................................... 16 Moisture .................................................................................................................. 16 Particle Size ............................................................................................................ 18 Temperature ............................................................................................................ 19 COMPOSITION ANALYSIS ....................................................................................... 20 SUMMARY................................................................................................................... 21 REFERENCES .............................................................................................................. 22 CHAPTER 2 ? EFFECT OF MOISTURE ON THE PHYSICAL PROPERTIES OF SWITCHGRASS PELLETS............................................................................................. 28 INTRODUCTION......................................................................................................... 28 MATERIALS AND METHODS .................................................................................. 30 Pelleting...................................................................................................................... 30 Composition Analysis ................................................................................................ 32 Structural Carbohydrate Analysis........................................................................... 32 Acid Insoluble Lignin ............................................................................................. 33 Acid soluble Lignin................................................................................................. 34 Ash .......................................................................................................................... 35 Particle size Distribution ............................................................................................ 36 Moisture Adjustment.................................................................................................. 37 Size............................................................................................................................. 37 Particle Density .......................................................................................................... 38 Bulk Density............................................................................................................... 38 Porosity....................................................................................................................... 39 Durability ................................................................................................................... 39 Hardness..................................................................................................................... 40 Moisture Sorption....................................................................................................... 40 Moisture Sorption Isotherms...................................................................................... 42 Data Analysis ............................................................................................................. 43 x RESULTS AND DISCUSSION.................................................................................... 44 Compositional Analysis ............................................................................................. 44 Particle Size Distribution ........................................................................................... 45 Size............................................................................................................................. 46 Particle Density .......................................................................................................... 47 Bulk Density............................................................................................................... 48 Porosity....................................................................................................................... 49 Durability ................................................................................................................... 50 Hardness..................................................................................................................... 51 Moisture Sorption Rate .............................................................................................. 53 Moisture Sorption Isotherms...................................................................................... 57 CONCLUSION ............................................................................................................. 63 REFERENCES .............................................................................................................. 65 CHAPTER 3 ? COMPACTION BEHAVIOR OF SWITCHGRASS.............................. 69 INTRODUCTION......................................................................................................... 69 MATERIALS AND METHODS .................................................................................. 71 RESULTS AND DISCUSSION.................................................................................... 75 Density ....................................................................................................................... 75 Effect of Pressure and Grind Size .............................................................................. 79 Specific Energy .......................................................................................................... 82 CONCLUSION ............................................................................................................. 85 REFERENCES .............................................................................................................. 86 CHAPTER 4 ? FUTURE WORK..................................................................................... 88 COMPACTION BEHAVIOR ....................................................................................... 88 SIMULTANEOUS SACCHARIFICATION AND FERMENTATION (SSF) ............ 89 ENZYMATIC HYDROLYSIS ..................................................................................... 91 PRETREATMENT........................................................................................................ 91 REFERENCES .............................................................................................................. 95 CONCLUSION................................................................................................................. 98 REFERENCES ............................................................................................................... 101 APPENDICES ................................................................................................................ 110 APPENDIX A.............................................................................................................. 111 APPENDIX B.............................................................................................................. 114 xi Figure 2.2 Pilot scale pellet mill??????????????????. 31 Figure 2.5 Diagram of environmental chamber used for moisture Figure 2.6 Particle size distribution of switchgrass grind at various switchgrass????????????????????....... 47 rass pellets???????????????????????? 48 Figure 2.9 and bulk density of switchgrass pellets??????????? 50 ??? 2 Figure 2.11 Effect of moisture content on durability and hardness of Figure 2.12 Moisture change in switchgrass pellets exposed to air at ................... 4 t 50% relative humidity and air at 15, 25, 35, and 45 ?C??????? 55 Figure 2.14 6, 20, 35, and 50 ?C?????????????????...... 58 ?? 2 Figure 3.2 Force-time curve for switchgrass made in a single pellet apparatus 74 Figure 3.3 switchgrass?????????????????????? 74 ntent (w.b.) during conditioning???????????????? 75 Figure 3.5 Figure 3.6 Density of switchgrass pellets with constant applied pressure ..... 0 Figure 3.7 Density of switchgrass pellets as particle size increases???...... 81 Figure 3.8 apparatus at constant die sizes??????????????.. 84 LIST OF FIGURES Figure 2.1 Hammer mill?????????????????????. 31 Figure 2.3 Figure 2.4 Ground and pelleted switchgrass?????????????.. Durability tester???????????????????? 32 40 sorption studies????????????????????. 42 sieve apertures????????????????????.. Effect of moisture content on length and diameter of 46 Figure 2.7 Figure 2.8 Effect of moisture content on particle density of switchg Effect of moisture content on porosity, particle density, Figure 2.10 Relationship between durability and particle and bulk density of switchgrass pellets??????????????? 5 switchgrass pellets??????????????????? 53 15?C and relative humidities of 50, 65, and 80%.......... Moisture change in switchgrass pellets exposed to air a 5 Figure 2.13 Moisture sorption isotherms for switchgrass pellets at Figure 2.15 Figure 3.1 EMC/ERH data fitted to four isotherm models?????? Single pellet compaction test set up?????...??????. 6 73 Force-distance curve used to calculate energy needed to compress Figure 3.4 Temperature profiles of switchgrass at 13.2% moisture co Density of switchgrass pellets affected by temperature and die size at constant moisture content of 10.4% and 20.0% (w.b.)??.. 78 of 95MPa and 10.4% moisture content (w.b.)???????. 8 Effect of temperature, moisture content and die size on specific energy required to make switchgrass pellets in a single pellet xii Table 2.1 Table 2.2 Particle size distribution of ground switchgrass????????.. 45 rium moisture content (Mq) obtained from non-linear regression analysis . Table 2.4 Model coefficients and values for the mean relative deviation 2 .. Table 3.1 switchgrass samples from three hammer mill screen ?????. 81 LIST OF TABLES Composition of ground and pelleted switchgrass???????? 44 Table 2.3 Estimated values of sorption rate constant (k) and equilib for switchgrass pellets?????????????????? . 56 (MRD) standard error of estimate (SEE) and coefficient of determination (R ) for a temperature range of 6-50?C??????????? Geometric mean diameter and standard deviation of ground . 61 sizes???????????????????? . 1 re on fos il fuel. The dep vulnerability to supply disruptions and to price volatilities. In addition, fossil fuels are non-renewable and there are environmental problems associated with the extraction, energy source One c switchgrass ? a high yielding peren cellent conservation attributes with a relatively high energy value - about 8000 BTU/lb (McLaughlin et al., 1999; M a b a cti n to here it can u to increase the bulk density of grasses such as alfalfa. Pelleting decreases transportation and storage co m r handling processing and storage equi Switchgrass is hygroscopic by nature. It will therefore exchange moisture with the surrounding environment. Knowledge of the moisture exchange rate for pellets under various conditions (temperature and relative humidity) is important for product integrity. INTRODUCTION The is a growing need in the United States to decrease dependence s endence on fossil fuel compromises the nation?s security in terms of transportation and utilization of fossil fuels. Energy from biomass is an alternative that can replace some the energy obtained from fossil fuels. rop that has been identified to have potential as a bioenergy crop is nial grass that has ex cLaughlin nd Kszos, 2005). Similar to other perennial grasses, switchgrass has a low ulk density nd therefore cannot be economically transported from place of produ o w be effectively utilized. Densification, mostly by pelleting, has been sed sts and makes the biomass easier to handle. Characterization of the echanical p operties of these pellets is required for design and sele pment and facilities. ction of , 2 The quality of pellets obtained from a pellet mill is affected by parameters such as moisture content, temperature, particle size, die size and pellet mill speed. Therefore, a study of the effect of these parameters on pellet quality during pelleting (at pilot scale or ta for the design of biomass handling and ese ompacted industrial level) is needed to provide da processing facilities. Compaction studies are a good way to investigate the effect of th parameters on the densification of switchgrass and can be carried out by means of a single-pellet apparatus system. Therefore, the goal of this project was to study the compaction behavior of switchgrass. This goal was achieved by the following specific objectives: 1. Quantify the effect of moisture content on the physical properties of c (pelleted) switchgrass; 2. Evaluate the effect of process parameters on the compaction behavior (density and specific energy) of switchgrass and, 3. Evaluate the effect of pelleting on the composition of switchgrass. 3 CHAPTER 1 ?LITERATURE REVIEW t rising oil prices are creating interest in the development l., ncept with eno global scales (Vinterback, 2004). Recent advances in biomass feedstock development and as a mean rger quantities instead of relying on woo ergy from biomass (i.e. bioenergy) has the potential to reduce dependency on fossil fuels thereby reducing the emission of greenhouse gases to the environment (Jannasch et al., 2001c; Mani et al., 2004a; Parikka, 2004). In addition, if bioenergy crops are grown in ecologically appropriate locations, their incorporation into agricultural systems could possibly provide extensive grassland bird habitat and address soil and water quality concerns (Roth et al., 2005). Bioenergy in the form of heat, electricity, and liquid fuels represents 14% of the world?s primary energy supply. About 25% of bioenergy usage occurs in industrialized countries with the remaining 75% used in developing countries. A comparison between BIOENERGY Increasing concerns abou of economical and convenient renewable energy fuels such as biomass (Samson et a 2000). The utilization of energy crops as a source of renewable fuels is a co rmous relevance to the current ecological and economic issues at both national and conversion technologies have created new opportunities for using agricultural land s of producing these renewable fuels in la d and agricultural residues alone (Samson et al., 2000). Renewable en 4 the available potential with current use indicates that on a worldwide scale two-fifths of the existing biomass potential is used which leads to the conclusion that current biomass use is below the available potential in most of the world (Parikka, 2004). In the United about 6%. A recent report by the U. S. Department of nergy indicates the country has the potential of generating over a billion tons of biomass annually for bioenergy use. The energ nt of biomass is sufficient to replace l, ng e t on ous and winter months. For herbaceous feedstocks, unprote States, bioenergy production is E y from this amou more than 30 % of the current petroleum consumption in the country (Perlack et al., 2005). Biomass stores energy during the process of photosynthesis. This energy can be recovered by the combustion process or by conversion into usable forms such as ethano bio-oils, producer gases, and pellets/briquettes (Mani et al., 2003a; Mani et al., 2004a). The net energy available from biomass ranges from 20 MJ/kg for dry plant matter to 55MJ/kg for methane, compared with about 27 MJ/kg for coal (Mani et al., 2003a). Biomass co-firing with coal is another way of utilizing renewable technology. Co-firi entails combusting in an existing coal-fired unit, a combination of biomass and coal. Th use of existing facilities reduces the capital investment and subsequently, the potential cost of the resulting renewable energy. Feedstock quality has a significant impac determining the optimal conversion processes. One requirement of large-scale biofuel production is the ability to store biomass feedstock for 6-12 months to ensure continu availability during off-season cted outside storage can result in losses that have a negative impact on the economics of the conversion process. Losses include decrease in mass and changes in composition of structural and non-structural components and arise from weathering and 5 al SWITCHGRASS Switchgrass (Panicum virgatum L.) is viewed as a major future energy crop in the United States (McLaughlin et al., 1999; Mani et al., 2004a), Canada (Samson et al., 2000; Christian et al., 2002) and Europe (Sharma et al., 2003). In the1980?s the Bioenergy Feedstock Development Program (BFDP) at Oak Ridge National Laboratory (ORNL) started screening more than 30 herbaceous crop species with the goal of developing a renewable energy source that can be used to produce transportation fuel and to generate electricity (Sanderson et al., 1996). In 1991, a decision was made by the ORNL to focus future work on switchgrass. This was because switchgrass is a high yielding, tall grass prairie species with relatively modest ash levels that has excellent conservation attributes and good compatibility with conventional farming practices (McLaughlin et al., 1999; Samson et al., 2000). Switchgrass also has a high fiber content and high biomass yield (Roth et al., 2005) and has good quality for heat and electricity production through thermal conversion (Sharma et al., 2003). Switchgrass tolerates diverse growing conditions, ranging from arid sites in biochemical reactions (Wiselogel et al., 1996). In addition to storage losses there are unavoidable losses of dry matter during field operations (e.g. cutting baling, transport) and during field curing of the plant material (Sanderson et al., 1997). Effective utilization of lignocellulosic feedstock can be deemed impractical because of its season availability, scattered stations and the high costs of transportation of large amounts of organic matter (Szczodrak and Fiedurek, 1996). 6 be used for bedding under animals, r the mushroom industry and for paper pulp production to replace hardwoods (Sharma et al., 2003). Due to the high productivity of the grass it can be grown by farmers on arginal land offering a cash crop and a boost to the farm economy (Boylan et al., 2000; Alizadeh et al., 2005). Zhan et al (2005) examined the feasibility of producing ethanol using switchgrass as the primary raw material. They found that while switchgrass is not currently grown commercially as a feedstock for energy production, it appears viable in the Southeast, Midwest, and Plains states. To support ethanol production, switchgrass can be grown at regional farms, cut, field dried, baled and transported by truck to a switchgrass-to ? ethanol conversion facility. The bulky nature of switchgrass results in high transportation cost for the raw material (Zhan et al., 2005). Pelleting switchgrass raises the bulk density of the feedstock eliminating high transportation costs when enroute to a conversion facility. A study carried out by Jannasch et al (2001a) found that the calorific value of switchgrass pellets (19.01 MJ/kg) is similar to that of wood (19.60 MJ/kg). short grass prairie to brackish marshes and open woods (Sanderson et al., 1996; McLaughlin et al., 1999) and requires little fertilization and herbicide (Boylan et al., 2000). The ecological diversity of switchgrass can be credited to three principle characteristics which include its open pollinated reproductive mode, a very deep, well- developed rooting system, and efficient physiological metabolism (Sanderson et al., 1996; McLaughlin et al., 1999). The crop can also fo m 7 hysical e l Thinner pellets give rise to a more uniform combustion rate especia orce iffering sizes, s , and hardness are aggregated. The point of aggregation is called a crack tip where stresses will accumulate. When the local stress at the tip of a crack becomes PHYSICAL PROPERTIES Knowledge of physical properties of biological materials is important for handling, storage, and transportation purposes. Some of the moisture dependent p properties of biological materials are shape, size (diameter and length), bulk and particl densities, porosity, hardness and durability (Balasubramanian, 2001). These physica properties can be used to determine whether a pelleted biological material will maintain its integrity during handling and transportation from the place it is manufactured to the place of utilization (Thomas and van der Poel, 1996). The dimensions of pellets in terms of length and diameter are important for combustion processes. lly in small furnaces. The length of the pellets affects the fuel feeding qualities, where, shorter pellets allow for an easier continuous flow (Lehtikangas, 2001). Bulk density affects the transportation costs and efficiency of handling and storage, thus the need to densify most biomass (Lehikangas, 2001; Sokhansanj and Turhollow, 2004). Strength properties include hardness and durability. Hardness describes the f needed to rupture the pellet (Thomas and van der Poel, 1996; Lehikangas, 2001). Fragmentation or rupture of inhomogeneous materials usually occurs near or at the point of inhomogeneities due to local stresses and strains being higher near points of imperfection. Fragmentation of homogenous materials can be errors in crystalline structure or small holes in or below the surface. Within pellets, particles of d hapes 8 higher than the cohesive or adh o grow and fracture occurs (Thoma s is advantageous when transpo en r armer teady ed from s esive stresses the crack begins t s and van der Poel, 1996). Durability describes the amount of fines produced after being exposed to mechanical or pneumatic agitation (Thomas and van der Poel, 1996; Lehtikangas, 2001). Fines are formed when pellets are dropped from the conveyor down to a pile. The fine can accumulate under transport conveyors resulting in an explosion. There is an increased tendency of fines to absorb moisture, which can leave the biomass susceptible to microbial attack. Therefore, high durability of pellets rting pellets at the plant or to the end user (Lehikangas, 2001). Durability is considered high when above 80%, medium when between 70% and 80% and low wh below 70% (Tabil and Sokhansanj, 1996; Tabil and Sokhansanj, 1997; Adapa et al., 2003a). Pellets are more sensitive to shearing at the places where they are cut off afte leaving the die (Thomas and van der Poel, 1996). Improper cooling can increase the pellet?s sensitivity due to stresses in the pellet between the cooled outer layer and the w center (Thomas and van der Poel, 1996; Lehikangas, 2001). During the cooling process, the cooling air is used to take up moisture and heat from the pellets. In a s state situation, the same amount of water is transported through capillaries to the surface of the pellet. When the speed of air is increased more water and heat are remov the pellet surface than can be delivered by the capillaries. This leads to a brittle outer layer with different physical properties than the interior of the pellet. These difference create stress, which cause the outer layer to crack under less than optimal conditions. The cracks allow for more fine formation (Thomas and van der Poel, 1996). 9 (1994) investigated the effect of increasing moisture on pell h re t es Moisture Sorption Isotherms Biological materials are hygroscopic in nature. These materials therefore, have the ability to exchange moisture with the atmosphere (Singh, 2004). Knowledge of the equilibrium moisture content (EMC) - equilibrium relative humidity (ERH) relationship is essential to designing and optimizing post harvest operations such as storage, drying, aeration, handling, packaging, and processing of biological materials (Singh, 2004; Durakova et al., 2005; Pangano and Mascheroni, 2005; Erbas et al., 2005). The equilibrium moisture content is the moisture content at which a hygroscopic material is in equilibrium with a particular environment. The relative humidity of that environment is The effect of moisture on physical properties has been studied for different types of biological materials. Fasina et al ets made from alfalfa. The bulk density of the alfalfa pellets decreased wit increase in moisture content. This was because the increase in mass due to moisture gain was lower than the increase in volume. Similar results have been reported for other biological materials (Deshpande et al., 1993; Nimkar and Chattopadhyay, 2001; Balasubramanian, 2001; McMullen et al., 2005). When this occurs, the amount of storage space required per unit mass of material increases with increasing moistu content. The opposite can also occur in biological materials indicating the change in volume is less than the corresponding change in mass with increase in moisture, (Joshi e al., 1993; Chandrasekar and Viswanathan, 1999; Jha 1999). Bulk and particle densiti are indicators of the pelletability of a material. Higher initial values may lead to good quality pellets and easier pelletability (Jannasch et al., 2001c). 10 the equ f cts are and and quality where high moisture can result in swelling, disintegration, and prevention of application for thermo- hemical conversion (Singh, 2004). In drying, the difference between the product moisture content and the EMC is often used as a measure for the driving force (Jenkins, 1989). EMC data are also used for thermodynamic analysis of water sorption. Thermodynamic properties that can be obtained from equilibrium moisture studies include heat of sorption, free energy, and entropy. The heat of sorption is beneficial in estimating the heat requirement during drying and the state of absorbed water in the solid materials. The level of moisture at which the heat of sorption approaches the heat of vaporization of water is taken to be indicative of the amount of bound water in the material of interest. At moisture contents ilibrium relative humidity (ERH). A plot of the EMC-ERH at a particular temperature is the moisture sorption isotherm. The moisture isotherm curve is often used to determine the storage stability o biological materials when exposed to varying environmental conditions during transportation and storage (Erbas et al., 2005). For example, most alfalfa pellets and cubes produced in Canada are exported overseas. While in transport, these produ exposed to high humidity and low temperature conditions which cause moisture uptake. This increase in moisture during exposure to humid conditions reduces the durability storage stability of the alfalfa pellets and cubes (Fasina and Sokhansanj, 1992; Fasina Sokhansanj, 1993). EMC data can also determine the lower and upper limit up to which biomass should be dried. Information about the mechanism of water binding during sorption can also be attained. Limited amounts of moisture are beneficial as the steam generated causes steam gasification reaction leading to better gas c 11 ilable 93). ral and for s ch water activity. Models adopted as standard equatio 05). le ds higher than this level, water is free in the void spaces of the system and readily ava for microorganisms (Fasina and Sokhansanj, 19 More than 200 equations (theoretical, semi-empirical and empirical) have been developed to determine the relationship between the EMC and ERH of agricultu biological materials. However, none of these equations describe the sorption process the entire range of water activity (Soysal and Oztekin, 2001). The best known isotherm equations are Brunauer-Emmet-Tetler (BET), Langmuir, Halsey, Henderson, Chung- Pfost, Chen-Clayton, Iglesias-Chirife, and Guggenheim-Anderson-de-Boer (GAB). Soysal and Oztekin (2001) concluded that the BET equation is suitable for most materials, especially for hydrophilic polymers below water activity of 0.5. The Hasley equation is appropriate for hydrophilic polymers and rubbers, plastics, synthetic fiber and foods rich in soluble components. The Chung-Pfost equation is suitable for cereal and other field crops, while the Iglesias-Chirife equation has been suitable for foods ri in soluble components. The GAB equation is considered the most versatile model for various materials over a wide range of ns by the American Society for Agricultural and Biological Engineers (ASABE) (ASABE Standard D245.5) includes Modified Chung-Pfost, Modified Henderson, Modified Halsey, and Modified Oswin equations (ASABE, 2001; Durkova et al., 20 The Halsey equation was developed for high protein and oil content food products whi the Modified Henderson and Chung-Pfost equations have been suitable for starchy foo (Fasina and Sokhansanj, 1993; Soysal and Oztekin, 2001). These equations have been used to model the moisture sorption isotherms of various agricultural and biological materials such as alfalfa pellets (Fasina and Sokhansanj, 1993), cotton plant components 12 Durakova and Menkov, 2005) and semolin of ate t (Barker, 1996), 10 medicinal and aromatic plants (Soysal and Oztekin, 2001), amaranth grains (Pagano and Masheroni, 2005), chickpea flour ( a and farina (Erbas et al., 2005). Usually in these studies, the researchers determined the equation that best fit the experimental data for various biological materials. In general, one or more of the following five parameters have been used to quantify the goodness of fit of the equations to moisture isotherm data: the coefficient determination (R 2 ), the residual sum of squares (RSS), the standard error of the estim (SEE), the mean relative deviation (MRD) and the plots of residuals (Fasina and Sokhansanj, 1993; Barker, 1996; Soysal and Oztekin, 2001; Erbas et al., 2005; Durkova et al., 2005; Pagano and Mascheroni; 2005). If the model is acceptable the residuals should be random independent errors with a zero mean, constant variance, and arranged in a normal distribution (Pagano and Mascheroni, 2005). If the residuals show a clear pattern, the model is unacceptable. Because all the information is contained in the residuals, the analysis of residuals is a valuable tool for diagnosis (Fasina and Sokhansanj, 1993; Pagano and Mascheroni, 2005). Models can be found unsuitable by more than one statistical criterion. Pagano and Masheroni, (2005) found that at firs glance R 2 values were relatively high for all models (R 2 >0.9803) showing that all models can be considered valid. However, the value of R 2 is not by itself a solid or robust analysis index. For example in the same study by Pagano and Masheroni (2005), it was found that the modified Halsey equation gave higher values of RSS, SEE and MRD and the smallest values of R 2 . This suggests that the Modified Halsey equation is not the most appropriate model for describing the experimental data. In short, low values of R 2 , 13 Rate of Moisture Sorption In the previous section, it was mentioned that biological material will sorb or desorb moisture until reaching equilibrium with the environment. This section deals with the rate at which the moisture sorption process takes place. As expected, several investigators have reported that the moisture sorption rate is a function of temperature and relative humidity of the environment and the type of material (i.e. chemical composition and the physical form ? e.g. cubes versus pellets (Zink, 1935; Dexter et al., 1947; Fasina and Sokhansanj, 1992). Fasina and Sokhansanj (1992) concluded that the moisture absorption rate for alfalfa cubes and pellets was affected by temperature and relative humidity. Chhinnan (1984) in designing an experimental dryer for thin layer studies investigated drying models to find the most appropriate for the application of drying pecans. The exponential model was found to be one of the simplest models to describe moisture movement in solids. This model assumes negligible internal resistance which means there is no resistance to moisture movement from within the material to surface. It only considers the surface resistance implying all the resistance is concentrated in a layer at the surface of the material. Other researchers have been successful at employing this model to describe moisture sorption rates in biological materials (Fasina and Sokhansanj, 1992; McMullen et al., 2005). Other mathematical models used to predict moisture high values of RSS, SEE, and MRD, and clear patterns in residual plots are indicative of the inability of the model to explain the variation in experimental data (Barker, 1996; Pagano and Masheroni, 2005). 14 r l (Dadgar et al., 2004). Dadgar et al (2004) found that the Page and two-tem experimental model best predicted moisture adsorption data for eld peas. mass ty ilable. Densification involves using some form of mechanical pressure to redu g ers sorption rates include: the Peleg model (Dadgar et al., 2004; Mali et al., 2005), Fick?s diffusion model (Chhinnan, 1984; Dadgar et al., 2004; Bello et al., 2005), Page (Dadga et al., 2004) two term exponentia fi COMPACTION Biomass has relatively low bulk density, which causes problems during storage, handling and transportation for further processing. The bulk density of some bio range from around 40 kg/m 3 for loose straw and bagasse to 250 kg/m 3 for some wood residues (Mani et al., 2003a). Densification of biomass into pellets or briquettes increases the volumetric energy content, reduces transportation costs and makes a varie of applications ava ce the volume of biological matter (Bruhn, 1989; Mani et al., 2003a). The mechanism of particulate bonding further explains the densification process. During the first stage of compression, particles that are preheated through dry blending or wet granulation, rearrange themselves to form a closely packed mass. Durin this phase, the original particles retain most of their properties and energy is dissipated due to inter-particle and particle-to-wall friction. At high pressures, the particles are forced against each other even more and undergo elastic and plastic deformation, increasing inter-particle contact. When the particles approach each other closely, short range bonding forces like van der Waal?s forces, electrostatic forces and sorption lay 15 ed the density of the pellet approaches the true densities of the component ingredients. There is a close correlation between the increase in density and the increase in applied pressure in the early stages of compression but the rate of increase in density lls off rapidly as the density of the pressed material approaches the density of water (Mani et al., 2003a). The overall goal during pelleting is to create a more fluid process, where a lower friction coefficient is created between the die extrusion surface and the fiber. This can be done during conditioning step that has been proven to improve pellet durability and production rates, decreasing the energy consumption of the pellet mill where steam commonly used for conditioning acts as a lubricant reducing friction during pelleting (Gilpin et al., 2002). The pellet is bound together by the lignin component of the feedstock. This process results when fiber passes through the extrusion holes, heating up the die and creating elevated temperatures. Lignin within the material starts to flow from the fiber cell walls and has the effect of binding with other fibers during extrusion (Jannasch et al., 2001a). During the process some moisture is driven off as stream. The resulting product is a uniform flowing material with a bulk density 2-4 times higher than that of the starting raw material. Since the composition of biological materials are not the same, material quality and type have been found to affect the compaction (or become effective (Chin and Siddiqui, 2000; Mani et al., 2003a). Under stress, brittle particles may fracture leading to mechanical interlocking. Mechanical interlocking is the only bonding mechanism that does not involve atomic forces and therefore contributes very little to the overall strength of the pellet. At higher pressures, the volume is reduc further until fa 16 oistu densification) of agricultural materials (Tabil and Sokhansanj, 1997; Samson et al., 2000). Process Variables Affecting the Compaction Process The properties of a biological material that mostly contribute to the amount of densification that can be achieved in a material are moisture content of material, grind characteristics and conditioning temperature (Jannasch et al., 2001a; Mani et al., 2003a). The effect of these properties on the densification efficiency is described next. M re Moisture content plays a large role in determining density and strength of densified biomass. The production of stable and durable pellets or briquettes require material to be conditioned to an optimum moisture content before pelleting (Wamuk and Jenkins, 1995; Mani et al., 2003a). Density is influenced by moisture in two ways First, by the change in mass with changing moisture content and second, by the ch volume of the particles as the moisture content changes below the fiber saturation point. The fiber saturation point is the moisture content of the material at which the cell walls are completely saturated while the cavities are liquid free (Jenkins, 1989). In the pelleting/briquetting process, moisture acts as a film type binder that strengthens the bonds between particles. In the case of organic and cellular products, moisture helps in promoting bonding by van der Waal?s forces by increasing the true area of contact of the particles (Lehtikangas, 2001). The right amount of moisture therefore, helps i s the onya . ange in n developing the self-bonding properties in lignocellulosic substances especially at elevated 17 nes. r, a high moisture content of the raw material can make the feedstock slippery. This can make the material slide easily through the holes of the die in the pellet mill 2001) concluded that when s the e studies at have been reported on the effect of moisture on biomass densification have been on briquettes. isture s th and . t basis). temperatures and pressures that are typically used in briquetting and pelleting machi Howeve thereby reducing pellet quality. Similarly, Lehtikangas ( materials are too dry they clog the die because the resistance rol the die holes exceed roller force. When this occurs, the pelleting operation stops and significant down time results due to the unclogging of the dies Therefore, the optimum moisture content that will give high quality pellets varies according to the type of biomass. Most of th th carried out Mani et al (2003a) found that the density of hay briquettes decreased as mo content increased from 28 to 44%. The authors found that when the feed moisture content was between 8 and 10%, the briquettes were strong and free of cracks and the briquetting process was smooth. At higher moisture levels (> 10%), the briquetting process was erratic and the briquettes were weak. Chin and Siddiqui (2000) investigated the characteristics of biomass including sawdust, rice husks, peanut shells, coconut fibers, and palm fruit fiber. The biomass wa densified at 5-7 MPa and tested to evaluate relaxation behavior, mechanical streng burning characteristics. The moisture content prior to compaction ranged from 5-30% Briquettes with moisture content of approximately 20% showed the least relaxation. Wamukonya and Jenkins (1995) produced relatively high quality briquettes from wood residue and wheat straw with an optimum moisture content range of 12-20% (we Since the wheat straw had lower moisture content, the durability of the manufactured 18 ture sic Particle Size briquettes was lower than that of the wood residue briquettes. In the feed industry, the general consensus is that high levels of heat and moisture in the conditioner achieve proper pelleting for grain-based diets high in starch (Chin and Siddiqui, 2000). Mois is important because the right amount develops self-bonding properties in lignocellulo substances at the elevated temperatures and pressures prevalent in briquetting and pelleting machines. The size (average and distribution) of particles is very crucial to the quality of pellets obtained during densification (Lehtikangas, 2001; Mani et al., 2003a; Adapa et al., 2004). Because of the nature of biomass, it has to be ground (usually by means of a hammer mill) to achieve the particle size that is optimum for pellet production. Reduction of particle size increases the total surface area, pore size of material, and the number of contact points for inter-particle bonding in the compaction process. Particle size of grinds will therefore have an effect on the quality of the densified masses. Grinding of biomass was found to contribute to the production of higher density briquettes with better durability, lower water absorption rates, and an increase in binding properties of the feedstock (Jannasch et al., 2001b; Lehtikangas, 2001; Mani et al., 2003a). It was reported, however, that wider particle size distribution is more suitable for compaction because the smaller (fine) particles can rearrange and fill the void spaces of larger (coarse) particles thereby producing denser and more durable compacts. Coarsely ground materials tend to give less durable pellets because they may create natural fissures 19 d e the Mani et al., 2004a). Similarly, it was found by Jannasch et al (2001b) that a reduction in screen size from 3.2 mm to 2.8mm for the fine grinding process for switchgrass appeared to produce an increase in pellet hardness. Adapa et al 004) reported results that indicated a decrease in hammer mill screen size (3.20 mm to1.98 mm) gave rise to higher durability of fractionated sun-cured and dehydrated alfalfa pellets. In agricultural and biological process engineering, the size (average and distribution) of particles is often determined according to the ASABE Standard S319.3 (ASABE, 2003). Projector and image analysis, gravitational sedimentation, centrifugal sedimentation and scanning electron microscopy (SEM) have also been used for determining particle size distribution (Yang et al., 1996). . in the pellet, which are susceptible to breakage. A combination of fine and medium grin is therefore essential for compaction (Mani et al., 2004a). In a study of 34 feed mills in the United States, it was reported that pellet quality appeared to be the highest in feed mills equipped with hammer mills having the 3.2-4.0 mm split screens. Plants using a hammer mill with 4.0 mm screen appeared to hav lowest pellet quality ( (2 Temperature Temperature plays a major role in stability and durability of the product, and in the amount of energy required for compaction. The addition of heat to the densification system can be by means of wet steam, preheating feed materials or using heated die apart from the frictional heat generated due to compression (Mani et al., 2003a). In the case of 20 khansanj, (1996) conducted a study f aking ng process to include conditioning of the ground switchgrass to at least 75?C before COMPOSITION ANALYSIS It must be noted that the composition of switchgrass can vary according to the time of year it was harvested and how it was stored (Dale et al., 1996; Wiselogel et al., 1996). It has been reported that the cellulose and lignin content of switchgrass increases es during the growing season, while the hemicellulose and other soluble compon e alfalfa, the addition of high temperature steam enhanced pellet durability and reduced energy consumption in the pelleting process. Tabil and So or improving the physical quality of alfalfa pellets by controlling and optimizing manufacturing parameters and found that pellet durability increased when the conditioning temperature was raised from 65 to 95?C. The conditioning step is very important because the overall goal is to create a more fluid pelleting process, by minimizing the friction coefficient between the die extrusion surface and the fiber m the fiber more pliable. The combination of these effects results in optimal extrusion of the feedstock through the dies. Since switchgrass is made up of at least 20% lignin Wiselogel et al., 1996; Esteghlalian et al., 1997; Alizadeh et al., 2005), it is necessary for the pelleti extrusion through the die (Samson et al., 2000). as the plant ag ents decrease (Dale et al., 1996). Changes in any feedstock constituents can hav negative effects on the profitability of both thermochemical and biochemical conversion processes (Wiselogel et al., 1996). 21 n in can ent gel et re gy ent techniques for ethanol production. Components of interest include: structural carbohydrates, ash and lignin. Ash i biomass feedstocks promotes char formation during pyrolysis, and forms fusible ash at high temperatures in combustion units used for steam generation. Lignin is the component that has the highest carbon and energy component. Changes in lign impact thermochemical and biochemical conversion processes. Increased lignin cont upon storage can be expected at the expense of carbohydrate components (Wiselo al., 1996). In this study the heat treatment during conditioning of the feedstock may give rise to higher lignin content in the pelleted sample. However in biochemical processes, the loss of lignin is viewed as a pretreatment to make the structural carbohydrates mo susceptible to hydrolysis (Saddler et al., 1993; Dale et al., 1996; Szczodrak and Fiedurek, 1996; Lynd, 1996; Wiselogel et al., 1996; Esteghlalian et al., 1997; Soderstrom et al., 2003; Liu and Wyman, 2005; Kim and Lee, 2005a; Kim and Lee, 2006). SUMMARY The potential for switchgrass to be used widely as an alternative to current ener resources is very promising. The densification of this important, lightly dense, biomass can be optimized by investigating processing parameters which in turn can lead to lower transportation costs and easier storage and handling. Understanding the effect of moisture on the physical properties of these densified masses is also important for maintaining quality. Knowing the composition of the feedstock is valuable for optimizing conversion and pretreatm 22 Adapa, P. K., Schoenau, G. J., Tabil, L. G., Sokhansanj, S., Crerar B.J. (2003). Pelleting Society of Agricultural Engineers Annual international Meeting, July 27-30, Las Vegas, characteristics of fractionated sun-cured and dehydrated alfalfa grinds, Applied switchgrass by ammonia fiber explosion (AFEX), Applied Biochemistry and Standard D245.5, St. Joseph, MI. ASABE (2002). Cubes, pellets, and crumbles--definitions and methods for determining sieving, ASABE Standard, S319.3, St. Joseph, MI. Balasubramanian, D. (2001). Physical properties of raw cashew nut, Journal of Agricultural Engineering Research, 78, 291-297. Barker, G. L. (1996). Equilibrium moisture content of cotton plant components, Journal of Agriculture Engineering Research, 6 Bello, M gelatinization in whole rice grain during soaking, LWT ? Food Science and Technology, evaluation, Biomass and Bioenergy, 19, 411-417. Bruhn, H.D. Mechanical treatment, In Biomass and Handbook, ed. Kitani,O. and Hall., offee, Journal of Agriculture Engineering Research 73, 227-234. Chhinnan, M. S. (1984). Evaluation of selected mathematical models for describing thin yer drying of in-shell pecans, Transactions of the ASABE, 27, 610-615. REFERENCES of fractionated alfalfa products. ASABE Paper No. 036069: pages 1-11, American Nevada, USA. Adapa, P. K., Tabil, L. G., Schoenau, G. J., Sokhansanj, S. (2004). Pelleting Engineering in Agriculture, 20, 813-820. Alizadeh, H., Teymouri, F., Gilbert, T. I., Dale, B. E. (2005). Pretreatment of Biotechnology, 121-124, 1133-1142. ASABE (2001). Moisture relationship of plant based agricultural products, ASABE density, durability, ASABE Standard, S269.4, St. Joseph, MI. ASABE (2003). Method of determining and expressing fineness of feed materials by 3, 353-364. . O., Tolaba, M. P., Suarez, C. (2005). Water absorption and starch In Press, Corrected Proof, Available online November 2005. Boylan, D., Bush, V., Bransby, D. I. (2000). Switchgrass co-firing: pilot scale and filed C.W., Gordon and Breach Science, Amsterdam, 1989, 452-459. Chandrasekar, V., Viswanathan, R. (1999). Physical and thermal properties of c la 23 Chin, O. C., Siddiqui, K. M. (2000). Characteristics of some biomass briquettes prepared coastal panic grass grown as a biofuel in Southern England, echnology, 83, 115-124. a during storage, An ASABE/CSAE Meeting resentation, Paper NO. 046089, ltural Engineers/Canadien ttawa, ale, B. E., Leong, C. K., Pham, T. K., Esquivel, V. M., Rios, I., Latimer, V. M. (1996). s, 16. 5-539. 8. ling and 56, 51-63. fa ASABE, Kansas City, MO, ne 19-22. under modest die pressures, Biomass and Bioenergy, 18, 223-228. Christian, D. G., Riche, A. B., Yates, N. E. (2002). The yield and composition of switchgrass and Bioresource T Dadgar, S., Tabil, L. G., Crerar, B. (2004). Moisture Adsorption of files peas as function of temperature and relative humidity American Society of AgricuP Society of Agricultural engineers, Annual International Meeting, August 1-4, O Ontario, Canada. D Hydrolysis of lignocellulosics at low enzyme levels: application of the AFEX proces Bioresource Technology, 56, 111-1 Deshpande, S. D., Bal, S., Ojha, T. P. (1993). Physical properties of soybean, Journal of Agricultural Engineering Research, 56, 89-98. Dexter, S. T., Sheldon, W. H., Waldron, D. I. (1947). Equilibrium Moisture content of alfalfa hay, Agricultural Engineering, 28, 295-296. Durkova, A. G., N.D. Menkov N. D. (2005). Moisture sorption characteristics of chickpea flour, Journal of Food Engineering, 68, 53 Erbas, M., Ertugay, M. F., Certel, M. (2005). Moisture absorption behavior of semolina and farina, Journal of Food Engineering, 96, 191-19 Estghlalian, A., Hashimoto, A. G., Fenske, J. J., Penner, M. H. (1997). Mode optimization of the dilute sulfuric a id pretreatment of corn stover, poplar and switchgrass, Bioresource Technology, 59, 129-136. Fasina, O.O., Sokhansanj, S. (1992). Hygroscopic moisture absorption by alfalfa cubes and pellets, Transactions of the ASABE, 35, 1615-1619. Fasina, O.O., Sokhansanj, S., (1993), Equilibrium moisture relations and hat of sorption of alfalfa pellets, Journal of Agricultural Engineering Research. Fasina, O.O., Sokhansanj, S., Irudayaraj, J. (1994). Thermo-physical properties of alfal pellets, Paper No. 946022, Annual International Meeting of Ju 24 nts in the pelleting rocess, Applied Engineering in Agriculture, 18, 331-338. rass biofuel pellets, Website: http://reap?canada.com Gilpin, A. S., Herrman, T. J., Behnke, K. C., Fairchild, F. J. (2002). Feed moisture, retention time, and steam as quality and energy utilization determina p Jannasch, R., Samson, R., de Maio, A., Adams, T. (2001a). Changing the energy climate: clean and green heat from g , ccessed on August 27, 2004. a Jannasch, R., Quan, Y., Samson, R. (2001b). A process and energy analysis of pelletizing switchgrass, Website: http://reap-canada.com, accessed on August 26, 2004. Jannasch, R., Samson, R., de Maio, A., Adams, T., Ho Lem, C. (2001c ). Changing the nergy climate: clean and green heat from grass biofuel pellets, Website: http://reap-e canada.com, accessed on August 27, 2004. Jenkins, B.M., Physical Properties of Biomass, In Biomass Handbook, Editors: Kitani, ., Hall, C.W., Gordon and Breach Science, Amsterdam, 1989, 860 ? 891. erties of makhana, Journal of gricultural Engineering Research, 72, 145-150. ysical properties of pumpkins seeds, ournal Agricultural Engineering Research, 54, 219-229. on of corn stover by mmonia recycle percolation process, Bioresource Technology, 96, 2007-2013. ver by hot water and aqueous mmonia treatment, Bioresource Technology, 97, 224-232 ues and igestibility of cellulose, ioresource Technology, 96, 1978-1985. biomass: chnology, economics, the environment, and policy, Annual Review of Energy and the naka, L. S., Yamashita, F., Grossmann, M. V. E. (2005). Water sorption nd mechanical properties of cassava starch films and their relation to plasticizing effect, Carbohydrate Polymers, 60, 283-289. O Jha, S. N. (1999). Physical and hygroscopic prop A Joshi, D. C., Das, S. K., Mukherjee, R. K. (1993). Ph J Kim, T. H., Lee, Y. Y. (2005a). Pretreatment and fractionati a Kim, T. H., Lee, Y. Y. (2006). Fractionation of corn sto a Lehtikangas, P. (2001). Quality properties of pelletised sawdust, logging resid bark, Biomass and Bioenergy, 20, 351-360. Liu, C., Wyman, C.E. (2005). Partial flow of compressed hot water through corn stover to enhance hemicellulose sugar recovery and enzymatic d B Lynd, L. R. (1996). Overview and evaluation of fuel ethanol from cellulosic te Environment, 21, 403-465. Mali, S., Saka a 25 ani, S., Tabil, L. G., Sokhansanj, S. (2004a). Grinding performance and physical cLaughlin, S.B., Bouton, J., Bransby, D., Conger, B., Ocumpaugh, W., Parrish, D., y crop, ress; 1999, p. 282?299. Applied Engineering in Agriculture, 21, 645- 51. een gram, ournal of Agricultural Engineering Research, 80, 183-189. n isotherms for amaranth grains, ournal of Food Engineering, 67, 441-450. ss and Bioenergy, 27, 613- 20. ch, D. C. 005). Biomass as feedstock for a bioenergy and bioproducts industry: the technical sion.pdf Mani, S., Tabil, L. G., Sokhansanj, S. (2003a). An overview of compaction of biomass grinds, Powder Handling and Processing, 15, 160-168. M properties of what and barley straws, corn stover, and switchgrass, Biomass and Bioenergy, 27, 339-352. M Taliaferro, C., Vogel, K. Wullschleger, S. Developing switchgrass as a bioenerg In: Perspectives on new crops and new uses, Edited by Janick J., Alexandria, VA: ASHS P McMullen, J., Fasina, O.O., Wood, C. W., Feng, Y. (2005). Storage and handling characteristics of pellets from poultry litter, 6 Nimkar, P.M., Chattopadhyay, P. K. (2001). Some physical properties of gr J Pagano, A. M., Mascheroni, R. H. (2005). Sorptio J Parikka, M. (2004). Global biomass fuel resources, Bioma 6 Perlack, R. D., Wright, L. L., Turhollow, A. F., Grahm, R. L. Stokes, B. J., Erba (2 feasibility of a billion ton annual supply, Website: http://feedstockreview.ornl.gov/pdf/billion_ton_vi , accessed on ecember 15, 2005. ine, L., Undersander, D. J., Bartelt, G. A. 005). Grassland bird response to harvesting switchgrass as a biomass energy crop, m Pretreatment of Lignocellulosic residues, In ioconversion of Forest and Agricultural Plant Residues, Edited by Saddler, J.N. (1993), isdelle, M., Lapointe, C. (2000). Assessment of pelletized iofuels, Website: http://reap.ca/Reports/pelltaug2000.html D Roth, A. M., Sample, D. W., Ribic, C. A., Pa (2 Biomass and Bioenergy, 28, 490-498. Saddler, J.N., Ramos, L.P., Breuil, C. Stea B Biotechnology in Agriculture No. 9, 73-92. Samson, P., Duxbury, P., Dr b , accessed on August 16, 2004. 26 , Ocumpaugh, W. R., Hussey, . A., Read, J. C., Tischler, C. R. (1996). Switchgrass as a sustainable bioenergy crop, . P., Wiselogel, A. E. (1997). Biomass losses during harvest nd storage of switchgrass, Biomass and Bioenergy, 12, 107-114. nd oderstrom, J., Pilcher, L., Galbe, M., Zacchi, G. (2003). Two-step steam pretreatment nd Agriculture, 20, 495-499. ions for some medicinal and aromatic plants, Journal of Agriculture Engineering esearch. 78, 57-63. e, Transactions of the ASABE, 47, 343-349. k, J. (1996). Technology for conversion of lignocellulosic biomass ethanol, Biomass and Bioenergy, 10, 367-375. of griculture, 12, 345-350. emple, S. J., van Boxtel, A. J. B. (1999). Equilibrium moisture content of tea, Journal , M., van der Poel, A. F. B. (1996). Physical quality of pelleted animal feed 1. riteria for pellet quality, Animal Feed Science Technology, 61, 89-112. Sanderson, M. A., Reed, R. L., McLaughlin, S. D., Wullschleger, S. D., Conger, B. V., Parrish D. J., Wolf, D. D., Taliaferro, C., Hopkins, A. A. M Bioresource Technology, 56, 83-93. Sanderson, M. A., Egg, R a Sharma, N., Piscioneri, I., Pignatelli, V. (2003). An evaluation of biomass yield stability of switchgrass (Panicum virgatum L.) cultivars, Energy Conversion and Management, 44, 2953-2958. Singh, R.N. (2004). Equilibrium moisture content of biomass briquettes, Biomass a Bioenergy, 26, 251-253. S of softwood by dilute sulfuric acid impregnation for ethanol production, Biomass and Bioenergy, 24, 475-486. Sokhansanj, S., Turhollow A. F.(2004). Biomass densification ?cubing operations a costs for corn stover, Applied Engineering in Soysal, Y., Oztekin, S. (2001). Comparison of seven equilibrium moisture content equat R Sun, Y., Cheng, J .(2004). Enzymatic hydrolysis of rye straw and bermuda grass using cellulases supplemented with ?-glucosidas Szczodrak, J., Fiedure to Tabil, L. G., Sokhansanj, S. (1996). Process conditions affecting the physical quality alfalfa pellets, Applied Engineering in A Tabil, L. G., Sokhansanj S. (1997). Bulk properties of alfalfa grind in relation to its compaction characteristics, Applied Engineering in Agriculture, 13, 499-505. T of Agricultural Engineering Research, 74, 83-89. Thomas C 27 urability and relaxation of sawdust and wheat- raw briquettes as possible fuels for Kenya, Biomass and Bioenergy, 8, 175-179. ell, J. A., Sanderson, . A. (1996). Compositional changes during storage of large round switchgrass bales, nsanj, S., Crerar, W. J., Rohani, S. (1996). Size and shape related haracteristics of alfalfa grind, Canadian Agricultural Engineering, 38, 201-205. brium moistures of some hays, Agricultural Engineering, 16, 51-452. d g strategies for potential switchgrass ? to ? ethanol conversion a, Biomass and Bioenergy, 28, 295-306. Vinterback, J. (2004). Pellets 2002: the first world conference on pellets, Biomass and Bioenergy, 27, 513-520. Wamukonya, L., Jenkins, B. (1995). D st Wiselogel, A. E., Agblevor, F. A., Johnson, D. K., Deutch, S., Fenn M Bioresource Technology, 56, 103-109. Yang, W., Sokha c Zink, F. J. (1935). Equili 4 Zhan, B. F., Chen, X., Noon, C. E., Wu, G. (2005). A GIS-enabled comparison of fixe and discriminatory pricin facilities in Alabam 28 HAPTER 2 ? EFFECT OF MOISTURE ON THE PHYSICAL PROPERTIES OF orth America. The crop has been designated as an energy crop by the US al., 2005; McLaughlin and Kszos, 2005) and because it has excellent conservation attributes and good compatibility with conventional farming practices (McLaughlin et al., 1999). Similar to other forage crops, switchgrass is lightly-dense (< 150 kg/m 3 ) when harvested (Sokhansanj and Turhollow, 2004) and therefore cannot be efficiently and economically transported over long distances to areas where effective utilization can occur. Densification by pelletizing is one of the effective methods used to increase the value of agricultural and biological materials (Fasina and Sokhansanj, 1996; Barger, 2003). Pellets are manufactured by grinding, conditioning (application of heat and/or moisture) and forcing the ground sample through dies that range in diameter from 4 to 12 mm or even larger (Fasina and Sokhansanj, 1993). Physical properties of switchgrass pellets will be needed for the proper design and selection of systems and equipment to store, handle and transport the pellets (Mohsenin, 1986). Several published studies have C SWITCHGRASS PELLETS INTRODUCTION Switchgrass (Panicum virgatum) is a high yielding perennial grass species that is native to N Department of Energy because of its high biomass production from which renewable sources of fuel and electricity can be generated (Missaoui et 29 revealed that moisture content has a significant influence on the physical properties of biological materials such as switchgrass pellets (Nelson, 2002; McNeill et al., 2004; McMullen et al., 2005). Some common physic size (length and diameter) ulk and particle density, durability, hardness, and moisture exchange. Length and diameter are important for combusti inner (small diameter) pellets give ris (Thomas and van der Poel, 1 n be alled a crack tip where stresses will accumulate. When the local stress at the tip al properties of interest include: b on processes where th e to a more uniform combustion rate, especially in small furnaces. The length of the pellets affects the fuel feeding qualities, where, shorter pellets allow for an easier continuous flow (Lehtikangas, 2001). Bulk density affects the transportation costs and efficiency of handling and storage, thus the need to densify most biomass (Lehikangas, 2001; Sokhansanj and Turhollow, 2004). Hardness describes the force needed to rupture the pellet 996; Lehikangas, 2001). Fragmentation or rupture of inhomogeneous materials usually occurs near or at the point of inhomogeneities due to local stresses and strains being higher near points of imperfection. Fragmentation of homogenous materials ca errors in crystalline structure or small holes in or below the surface. Within pellets particles of differing sizes, shapes, and hardness are aggregated. The point of aggregation is c of a crack becomes higher than the cohesive or adhesive stresses the crack begins to grow and fracture occurs (Thomas and van der Poel, 1996). Durability describes the amount of fines produced after pelleted materials are exposed to mechanical or pneumatic agitation (Thomas and van der Poel, 1996; Lehtikangas, 2001). It is therefore a measure of the rate at which fines will be generated 30 the pellets (Aarseth and Prestlokken, 2003). This is because fines the s, 1989). High moisture content in storage may lea e collected from the E.V. Smith E , during the handling, storage and transportation of pelleted materials. It is generally desired that the integrity of pellets (i.e. minimal fines generation) be maintained until time of utilization of the have been known to contribute to explosion and fire hazards (Woodcock and Mason, 1987). In addition, fines from pelleted materials generally absorb moisture from the environment, which has been found to increase the susceptibility of the bulk material to microbial attack (Lehikangas, 2001). Biomass materials are hygroscopic and will therefore exchange moisture with environment. Knowledge of the rate of moisture uptake and equilibrium moisture relationships are important for characterizing product quality during storage and for defining optimal storage conditions (Jenkin d to loss in fuel values, growth of microorganisms and spontaneous combustion of biomass material intended for thermal conversion (Jenkins, 1989). The objective of this chapter is to investigate the effect of (a) pelleting on the composition of switchgrass and (b) increasing moisture on the physical properties of switchgrass pellets. The physical properties of interest include size (length and diameter), bulk and particle density, porosity, hardness and durability. MATERIALS AND METHODS Pelleting The switchgrass samples that were used in this study wer xperiment Station (Auburn University), Tallassee, Alabama. Before pelleting the switchgrass was ground through a 3.18 mm screen using a hammer mill (New Holland Grinder, Model 358, New Holland, PA) shown in Figure 2.1. A laboratory scale 31 d to e die size used was 4.76 mm. After pelleting, the pellet mill (Model CL5, California Pellet Mill Co., San Francisco, CA) was used to manufacture the pellets used in this study (Figure 2.2). The pellets were made by extruding the ground switchgrass through round cross- sectional dies. Before passing through the pellet die, the feed material was conditione by increasing the temperature to 75?C and increasing the moisture content from 15.6 30.1%. Frictional heating of the die during pelleting further increased the temperature of the pellets exiting the die to 85?C. Th pellets were cooled in an environmental chamber set at 22?C and 40% relative humidity. Figure 2.1 ? Hammer mill Figure 2.2 ? Pilot scale pellet mill Figure 2.3 ? Ground (3.18 mm screen) and pelleted switchgrass (4.76 mm diameter) Composition Analysis Composition analysis was performed on ground and pelleted switchgrass accord Procedure - Determination of Structural Carbohydrates an 2004). The pelleted and ground switchgrass rt Grind, Black and Decker Corp., Towson, MD). Structural Carbohydrate Analysis ing to National Renewable Energy Laboratory (NREL) Laboratory Analytical d Lignin in Biomass (NREL, was ground using a coffee grinder (Sma Structural carbohydrates (glucose, xylose were determined by using a two duplicate. In the first stage, each sample (groun in a con L f 72% sulfuric acid for 1 hour. After 1 hour the acid was diluted to 4% with 84 ? 0.04 g of dionized water using a balance (Model AE-160, Mettler Toledo Co., Columbus, OH) accurate to 0.01 g. In the second stage the diluted samples were autoclaved (Model 2540E, Tuttnauer Brinkmann, Ronkonkoma, NY) at 121 ?C for 1 hour. Sugar recovery , arabinose, galactose, and mannose) stage sulfuric acid hydrolysis. Samples were prepared in d and pelleted) was incubated at 30 ? 3?C stant temperature shaking bath (Model BT-25 Yamato, Lorton, VA) with 3.0 m o 32 33 standards (S to correct for losse dilute acid hydrolysis. The Samp onitored with pH paper). After neutralization, the inutes at 16,000g (17308-Series Micro-centrifuge nt Co., Vernon Hills, IL). HP tions ere as follows: Injection volume ? 10-50 ?L Flow rate ? 0.55 mL/minute Detector temperature ? close to column temperature Run time ? 35 minutes Acid Insoluble Lignin RS) were prepared and taken through the second stage. The SRS were used s due to the destruction of sugars during SRS included D-(+) glucose and D-(+) xylose. les were neutralized with calcium carbonate to pH 5-6 (this was m samples were centrifuged for ten m , Cole Parmer Instrume LC was used to quantify the amount of released sugars in the feedstock by a Bio-Rad Aminex HPX-87P column (Bio-Rad Laboratories, Hercules, CA). The HPLC condi w Mobile phase ? HPLC grade dionized water, .2 ?m and degassed Column temperature ? 80-85?C Detector ? refractive index To determine the acid insoluble lignin (AIL) content, the samples (from the two stage hydrolysis) were vacuum filtered through a filter crucible using dionized water to transfer all remaining solids. The filtrate was captured in a filtering flask and approximately 50 mL of filtrate was collected in a storage bottle to determine acid soluble lignin and carbohydrates. The crucibles containing acid insoluble residue were dried in an oven (ESP-400C, StabilTherm, Blue Mountain Electric Co., Blue Island, IL) at 105 ?3 ?C until a constant weight was achieved. After drying in the oven, crucibles were placed in a muffle furnace (Type-1500, Thermolyn Sybron Corp., Milwaukee, WI) 34 sing Equations 2.1 ? 2.3. at 575 ? 25?C for 24 ? 6 hours after which the crucibles with ash were cooled and constant weight was recorded. Samples were prepared in duplicate. Acid insoluble residue (AIR) and lignin (AIL) were calculated u 100 %TSWA ODW ? = (2.1) 100% ? ? = ODW CCR AIR (2.2) 100 )()( % ? ??? = CCACCR AIL (2.3) ible (g) CR = weight of crucible plus acid insoluble residue (g) WA = weight of air dried sample (g) TS = Total solids in sample (%) ODW Where, ODW = oven dried weight (g) C = weight of cruc CA = weight of crucible plus ash (g) Acid soluble Lignin To analyze for acid soluble lignin (ASL), the absorbance of the hydrolysis liquor was measured at a wavelength of 320 nm with a UV-Visible spectrophotometer (Model Synergy HT, Biotek Inc., Woburn, MA). ODW DilutionVUVabs ASL filtrate ? ?? = ? % (2.4) s dss V VV Dilution + = Where, UVabs = average UV-Vis absorbance for the sample at 320 nm V filtrate = volume of filtrate (87 mL) V V ds = volume of diluting solvent (mL) ? = absorbance of biomass at specific wavelength Ash (2.5) s = volume of sample (mL) For ash analysis, approximately 0.900 g of switchgrass (ground and pelleted) was placed in crucibles which were put in the aforementioned muffle furnace at 575 ? 25?C for 24 ? 6 hours according to NREL Laboratory Analytical Procedure - Determination of ash in biomass (NREL, 2005). After the allotted time, crucibles and ash were removed from the furnace and put directly into a dessicator and constant weight was recorded. sis was performed at the same time using a moisture analyzer Total solid (TS) analy (Model IR-30, Denver Instruments, Arvada, CO). The percent of ash in the sample was calculated from Equation 2.6: 35 ODW CCA Ash ? =% (2.6) 36 Particle size Distribution Particle size distribution was determined according to ASABE Standard S319.3 (ASABE, 2003). This involves placing 100 grams of material on the top sieve of a nest of successively smaller si ter the test is complete. For this analysis, 7 U.S. Series test sieves plus a pan with were used. The nest of test sieves was inutes in a Sieve shaker oved from the etric ard S319.3 (ASABE, 2003) (Equations 2.7 ? 2.9): eves and recording the weight of material retained on each sieve af aperture sizes ranging from 1.70 to 0.212 mm shaken for ten m (Model CL 340, Soil Test Engineering Test Equipment Co., Evanston, IL), after which, the nest of sieves were rem autoseiver unit and the mass of material retained on each sieve was recorded. The determination of the geometric mean diameter (d gw ) of the sample and the geom standard deviation of particle diameter (S gw ) was carried out according to the ASABE Stand ? ? ? ? ? ? ? ? ? ? n i ii gw W dW )log( ? ? ? ? = = =? i n i d 1 11 log (2.7) 2/1 1 1 ? ? ? ? ? ? = = = i i i 2 log )log(log ? ? ? ? ? ? ? ? ? n n gwii W ddW S (2.8) [ ( ) ] 1 11 2 1 ? ?? = dS ggw loglog loglog ? SS w (2.9) 37 ean diameter or median size of particles by mass (mm) The evaluation of the effect of moisture on the physical properties of the pellets was carried out at five moisture levels (6.3, 8.6, 11.0, 14.8, and 17.0%, (wet basis)). The initial moisture content of the pellets was 6.3% (wet basis). To adjust the moisture content of the pellets to the desired levels, the pellets were placed in an environmental chamber (Model AA-5460A, Espec Corp., Hudsonville, MI) set to 30?C and 90% relative humidity. The pellets were rem ber after gaining the desired amount of moisture. Conditione equilibrate for a period of 24 hours before testing. Before experimentation, the moisture content of the conditioned pellets was verified with a moisture analyzer (Model IR-200, Denver Instruments, Arvada, CO) Size ellets were weighed using a digital balance accurate to 0.001 grams (Model AR3130, Ohaus Corp., Pinebrook, NJ). The length (L (Model CD-56C, Mitutoyo Corp., Kawasaki, Japan) accurate to 0.01 mm. Where, d gw = geometric m S log = geometric standard deviation of log normal distribution by mass in ten bases logarithm (dimensionless) S gw = geometric standard deviation of particle diameter by mass (mm) W i = mass on i th sieve (g) n = number of sieves plus one pan d i = nominal sieve aperture size of the i th sieve (mm) Moisture Adjustment oved from the cham d pellets were then allowed to Fifty random pellets were selected for size evaluation. The p ) and diameter (D) were measured using a digital caliper 38 (Model AccuP ference the material. Based on the pressure difference the pycnometer calculates the volume of the ratio of the mass of p ) cell to the volume (V p ) measured by the pycnometer quation 2.10). Sample mass was obtained with a digital balance accurate to 0.001 grams (Model AR3130, Ohaus Corp., Pinebrook, NJ). This procedure was performed in triplicate and the average value was reported. Particle Density The particle density of the pellets was measured by gas comparison pycnometry yc 1330, Micromeritics Instrument Corp, Norcross, GA) where a known quantity of helium under pressure is allowed to flow from a previously known re volume into a sample cell containing between the sample cell and the reference cell, the material in the sample cell. Particle density (? p ) was taken as material in the sample (m (E p m ) material was leveled with the top surface of the container and weighed. The bulk density b ) of the pellets ?was taken as the mass of sample in the container (m c ) over the volume of the container (V c ) (Equation 2.11). This procedure was performed in duplicate. p p V =? (2.10 Bulk Density Bulk density was determined by a bulk density measurement apparatus (Burrows Co., Evanston, IL) and according to ASABE Standard S269.4 (2002). This method involves filling the container of the apparatus (volume of 947 mm 3 ) via a funnel. The (? c b V =? (2.1) c m Porosity 39 y Porosity is the percentage of the total container volume occupied by air spaces when particulate solids are placed in a container (Fasina and Sokhansanj, 1993). Porosit is mathematically defined by Equation 2.12 and was calculated using the average bulk and particle densities that were obtained in the previous sections. p b ? ? ? ?=1 (2.12) Durability The durability (Du) of the pellets was determined according to ASABE Standard S269.4 (2002). A 100 gram sample of pellets was tumbled at 50 rpm for 10 minutes, in a dust-tight enclosure (see Figure 2.4). A No. 5 U.S. Sieve with an aperture size of 4.0 mm was used to retain crumbled pellets af ratio of mass of pellets retained on the sieve after tumbling (m pa ) to mass of pellets before (Equation 2.13). Durability is said to be high when the computed value is above 8 ter tumbling. Durability is expressed by the percent tumbling (m pb ) 0%, medium when between 70% and 80%, and low when below 70% (Tabil and Sokhansanj, 1996; Tabil and Sokhansanj, 1997; Adapa et al., 2003). 100* pb pa m m Du = (2.13) 40 Figure 2.4 ? Durability Tester ardness software provided by the ma as a measure of pellet hardness (Thom Moisture Sorption An air-tight environmental chamber (1.8 m x 0.9 m x 0.9 m) was used to investigate the moisture sorption properties of the pellets. Air that was fed into the chamber was conditioned by a temperature-humidity conditioner (Model AA-5460A, H The hardness for the switchgrass pellets was determined using a texture analyzer (Model TA-HD, Stable Micro Systems, Surrey, UK). Fifty pellets that were randomly selected from each sample lot were used in this test. A single pellet was placed on the platform of the texture analyzer in its natural position (the radial dimension was in the same direction as that of the compressive force). A flat plate (50.8 mm diameter) plunger was pressed onto each pellet at a speed of 10 mm/s. The maximum force required to rupture the pellet was determined from the force-deformation curve recorded by the nufacturer. This force was taken as and van der Poel 1996; Adapa et al., 2003). 41 Parameter Generation and Contro binations of four air temperatures (15, 25, 35, and 45 idities (50, 65 and 80% (? 3%)) were used to i switchgrass pellets. To conduct of pellets (300-340 g) was placed in a wire me e accurate to 0.01 grams (Model, PM at ten minute intervals during exposure to the conditioned air by Software Ltd., Manchester, rface the balance to a personal computer. A data acquisition system (OMB- Daq-56 ure d average values were reported. l Inc., Black Mountain, NC). Com ?C (? 1 ?C)) and three air relative hum nvestigate moisture sorption properties of a test, a thin single layer sh basket that hung from a digital balanc 4600, Mettler-Toledo, Columbus, OH). The weight of the sample was monitored and recorded using Windmill RS232 communication software (Windmill UK) to inte , Omega Engineering Inc., Stamford, CT) was used to record and monitor the temperature and relative humidity of the conditioned air at ten minute intervals. Depending on the temperature and relative humidity of the conditioned air, the pellets took between 22 and 68 hours to equilibrate with the conditioned air. An experiment was stopped when the weight of the sample did not change by more than 0.01g within a span of one hour. A moisture analyzer (Model IR-200, Denver Instruments, Arvada, CO) was used to cross check the final moisture content of the pellets. The schematic diagram of the system is shown in Figure 2.5. Each temperat and relative humidity combination was performed in duplicate an 42 s Moisture Sorption Isotherms equilibrium moisture relationship of the switchgrass pellets. This method consists of bringing air into equilibrium with a material of fixed moisture content. This method has been found to be simpler and faster than bringing a sample to equilibrium with air at a fixed temperature and relative humidity (Fasina and Sokhansanj, 1992). The pellets were adjusted to desired moisture levels (4.5, 7.2, 9.0, 11.0, and 15.0%, wet basis) in an environmental chamber (Model AA-5460A, Espec Corp., Hudsonville, MI) set at 30?C and 90% relative humidity. Each sample was allowed to equilibrate for 24 hours before each test. The moisture content was verified using a moisture analyzer (Model IR-200, Denver Instruments, Arvada CO). Figure 2.5 ? Diagram of environmental chamber used for moisture sorption studie The equilibrium relative humidity (ERH) method was employed to determine the Switchgrass pellets Basket Conditioned air from Humidity chamber Air returned to Humidity chamber Balance Pc 43 A water activity instrument (HygroLab 2 - H3, Rotronic Instrument Corp., was continuously recorded on a personal computer until equilibrium was reached (usually less an 4 hrs). Water activity/moisture sorption analysis was carried out in triplicate. EMC and ERH were taken as the average of the three moisture contents and relative humidities for each sample. el Huntington, NY) was used to measure the equilibrium relative humidity of preconditioned samples. To conduct a test, the conditioned sample was placed in the sample holder of the water activity instrument. A sealed measurement system was formed by placing the water activity probe on top of the sample holder. The probe is equipped with a small fan that circulates air within the sample container, a thin film capacitance sensor that is capable of measuring relative humidity from 0 to 100% with an accuracy of ? 1.5% and a platinum RTD (resistance temperature detector) temperature probe with an accuracy of ?0.3?C. The measurement system was then transferred into the temperature controlled-chamber set at the desired temperature of 6, 20, 35 or 50?C. The relative humidity and dry bulb temperature output from the water activity meter th Data Analysis Statistical analysis was performed on all data sets using SAS statistical software package (Version 9.1, SAS Institute Inc., Cary, NC, 2002-2003) and Microsoft Exc (Windows XP, 2003). 44 Compositional Analysis arabinan and galactan from NREL laboratories. At first glance, the pelleted switchgrass had higher values than the ground samples for all properties evaluated except for ash. This could be due to the thermal and mechanical treatment which occurs during the pelleting process. Statistically there were no differences in carbohydrate levels, but there were significant differences in ash, acid soluble lignin and acid insoluble lignin at a 95% confidence level. However, there is no scientific explanation for the increase in the proportions of the variouscomponents of the sample due to pelleting. able 2.1 ? Composition of ground and pelleted switchgrass RESULTS AND DISCUSSION The composition of ground and pelleted switchgrass was similar (Table 2.1). Comparable results were reported by other researchers for the compositional analysis of ground switchgrass (Wiselogel et al., 1996; Esteghlalian et al., 1997; Alizadeh et al., 2005). Alizadeh et al (2005) reported results of 34.2% glucan, 22.1% xylan, and 3.1% T Component Ground Stdev Pellet Stdev AIL (%) 25.85 1.02 26.70 1.42 a ASL b (%) 13.02 0.75 14.93 0.01 Glucan (%) Xylan (%) 35.76 1.94 36.00 0.04 21.26 1.29 21.61 0.04 Galactan (%) 0.67 0.94 0.82 0.05 Arabinan (%) 2.21 0.47 2.39 0.05 Mannan (%) 0.99 0.15 0.84 0.06 Ash (%) 1.36 0.74 4.05 3.68 Note: a = Acid insoluble lignin, b = Acid soluble lignin 45 Particle Size Distribution The particle size distribution of the ground switchgrass that was used to make the ellets is shown in Figure 2.6 and Table 2.2. Most of the particles (29.51 and 38.55%) were retained on sieves with aperture sizes of 0.595 mm and 0.850 mm. The geometric mean diameter (d gw ) and the geometric standard deviation (S gw ) of the ground switchgrass were calculated to be 0.867 mm and 0.357 mm respectively (Equations 2.7- 2.9). This result, according to other researchers, indicates the particles are not highly compressible. Particles with sizes below 0.400 mm are considered fine and highly compressible. An increase in the amount of fine particles is usually associated with decreased flowability. The more compressible a powder is, the less flowable it will be and vice versa (Tabil and Sokhansanj, 1997; Mani et al., 2003a). Table 2.2 ? Particle size distribution of ground switchgrass Sieve Aperture Size (mm) Distribution (%) 12 1.700 10.75 p U.S. Sieve No. 20 0.850 38.55 40 0.425 7 6.28 2.07 1.55 4.29 30 0.595 29.51 7.50 50 0.29 60 0.250 70 0.212 pan 0 46 0 15 40 a D i b n (% 5 10 20 25 30 35 0.0 0.5 1.0 1.5 2.0 Sieve Aperature Size (mm) P r t ic le S i z e i s t r u tio ) Figure 2.6 ? Particle size distribution of switchgrass grind at various sieve apertures he effect of moisture content on the length (L) and diameter (D) of the pellets is illustrated in Fi he pellets, reaching a maximum of 35.27 mm at 8.62% moisture content. Further increase in moisture redu the length of the pell 3.61 mm. The diameter of the pellets varied from 4.85 mm to 5.25 mm and indicated an increase in diameter with increase in moisture content. The change in length and diameter of the pellets w e to water filling the void spaces of the pellets an ting bonds formed during the compaction process. The effects of moisture content the length and diameter of the pellets are given in the equations below: For length: 2 0943.07024.177.26 MML ?+= , R 2 = 0.872 (2.14) Size T gure 2.7. There was a slight initial increase in the length of t ced ets to 3 as du d disrup on 47 (2.15) Wh For diameter: 2 0052.00855.019.5 MMd +?= , R 2 = 0.978 ere, M is moisture content (%, wet basis). 20 30 35 40 0 5 10 15 20 e n gt h ( m m 4.74 4.84 5.04 5.14 5.24 5.34 5.44 i am et e r ( m m ) ) length diameter 25 L 4.94 D predicted Moisture Content (%, w.b.) In general, particle density of the pellets decreased with increase in moisture content (Figure 2.8). A maximum value of 1462 kg/m was obtained at moisture content of 8.62% (w.b.). The decrease in particle density is due to the expansion of the pellet, hence, an increase in the volume of the pellets with increase in moisture content. Similar trends were reported for other biological materials (Joshi et al., 1993; Deshpande, 1993). Equation 2.16 shows the effect of moisture content on the particle density of the pellets c ?+= , Figure 2.7 ? Effect of moisture content on length and diameter of switchgrass pellets Particle Density 3 yn p ? R 2 484.0435.98.1412 MM 2 = 0.940 (2.16) 48 1350 1400 1450 ( k g / 1500 0 5 10 15 20 Moisture content (%, w.b.) P a r t ic le d e n s ity m 3 ) Figure 2.8 - Effect of moisture content on particle density of switchgrass pellets Bulk Density The bulk density of the pellets for different moisture levels ranged from 53 ore switchgrass by more than three fold (the bulk density of ground switchgrass was ts slightly increased with moisture content equation describes the relationship between bulk density and moisture content: 6 kg/m 3 to 708 kg/m 3 . Pelleting therefore, reduced the amount of space required to st 169.5 kg/m 3 ). Initially, bulk density of the pelle , reaching a maximum of 708 kg/m 3 at 8.62% moisture content, and then decreased with further increase in moisture content (Figure 2.9). This is a result of the increase in mass due to moisture gain being lower than the accompanying volumetric expansion of the bulk. Therefore, the amount of storage space required for a given sample will increase with moisture addition. Similar results have been reported for other biological materials (Deshpande et al., 1993; Fasina and Sokhansanj, 1994; Nimkar and Chattopadhyay, 20 len et al., 2005). The following 01; Balasubramanian, 2001; McMul 49 (2.17) nsity e of inant orosity isture e was an initial decrease in porosity and a minimum of 51.61% at 8.62% oistur o ity r ther imkar 8) 2 2094.2518.3715.538 MM b ?+=? , R 2 = 0.988 Figure 2.9 compares the bulk and particle densities of the pellets. Bulk de displayed a higher sensitivity to the change in moisture content with a percent chang 24% compared to the percent change in particle density of 16%. Both showed a decrease as moisture content increased which is due to volumetric expansion being the dom effect. Because bulk and particle densities are indicators of the pelletability of the material, high values result in good quality pellets (Tabil and Sokhansanj, 1997). P The porosity of the pellets ranged from 51.61 to 62.62% with varying mo contents. Ther m e content. As moisture content continued to increase the porosity began t increase. Figure 2.9 and Equation 2.18 show the non-linear relationship between poros and moisture content. A comparison of porosity of switchgrass pellets with that of othe biological materials revealed that the effect of moisture was similar. However, the o biological materials had no minimum value and displayed a linear relationship (N and Chattopadhyay, 2001; Balasubramanian, 2001; McMullen et al., 2005). 2 1381.03015.2756.61 MM +?=? , R 2 = 0.990 (2.1 400 600 800 1000 1200 00 e n s i y ( k g/ m 3 ) 20 30 40 50 60 70 P o o s i ty (% ) 14 1600 0 5 10 15 20 Moisture content (%, w.b.) D t 80 r bulk density particle density porosity predicted Figu ulate bonds (Fasina and Sokhansanj, 1992). A two phase system of results for particles and water with no capillary force present to maintain the pellet structure. This leaves the pellet with cracks which makes the pellets susceptible to breakage (Thomas and van der Poel, 1996). According to the durability rating cited by Adapa et al (2003), the switchgrass pellets in this study can be classified to have high durability (89.06% to 95.91%) at re 2.9 ? Effect of moisture content on porosity and particle and bulk densities of switchgrass pellets Durability Durability (Du) of the pellets increased initially with moisture content and reached a maximum of 96.65% at 8.62% moisture content. Further increase in moisture content reduced durability (Figure 2.10) to 78.44%. Similar trends have been reported for alfalfa pellets (Fasina and Sokhansanj, 1992) and poultry litter pellets (McMullen et al., 2005). It is suspected that initially, the binding forces of the water molecules strengthened the bond between individual particles in the pellets. Further increase in moisture caused the disruption of the partic 50 moisture contents between 6.32% and 14.84%; and medium durability at moisture content values greater than 14.84%. Equation 2.19 shows the relationship between durability and the moisture content of the pellets 2 25.03179.4218.78 MMDu ?+= , R 2 = 0.976 (2.19) It can be noted that durability is highly correlated with bulk and particle density (Fig. 2.11) with correlation coefficients of 0.933 and 0.989 respectively Hardness Similar to durability, pellet hardness (H) generally decreased (30.21 N to 21.6 N) 51 wi eaving the pellets weak and susceptible to breakage. The lationship between pellet hardness and moisture content is shown in Equation 2.20. 2 ??= 2 Pellet hardness (H) displayed a higher sensitivity to moisture with a percent (McMullen et al., 2005). Several studies examined the effect of moisture on durability both properties were stable at moisture contents up to 10%, but decreased rapidly as ties is given in Table A.1 in Appendix A. th increase in moisture content (Figure 2.10). This is again, due to the moisture disrupting particulate bonds l re 024.02377.0852.32 MMH , R = 0.928 (2.20) change of 28.1% from 6 to 17% moisture compared to durability (Du) with a percent change of 18.2% also from 6 to 17% moisture as was seen with poultry litter pellets and hardness (Fasina and Sokhansanj, 1992; Khoshtaghaza et al., 1999) which found that moisture increased. In this study there was a rapid decrease in pellet hardness and durability after moisture content of 8.62%. A summary of all proper 52 35 40 ) 95 10 20 25 a r 80 u r a 30 s s ( 5 ilit y 0 5 10 15 20 H dne N 70 75 8 90 0 D b (% ) hardness durability predicted Moisture content (%, w.b.) Figure 2.10 ? Effect of moisture content on durability and hardness of switchgrass p ellets 60 70 80 90 100 500 550 600 650 700 750 Bulk density (kg/m 3 ) D u r a b ility ( % ) 60 70 80 100 1430 1440 1450 1460 1470 Particle density (kg/m 3 ) D u r a b ility ( 90 % ) Figure 2.11 ? Relationship between durability and particle and bulk density of switchgrass pellets Moisture Sorption Rate The typical effect of air temperature and relative humidity on rate of moisture absorption by the pellets is shown in Figures 2.12 and 2.13. In general, rate of moisture absorption varied with relative humidity and temperature of the air to which the pellets were exposed. A 15% increase in relative humidity (from 65 to 80%) at constant temperature of 15?C gave an increase of 3.25% in final moisture content. This was an 53 54 indication th e final moisture co se in relative hum A temp 45?C, the f eratures (Figure 2.13 mo at increased moisture in the air can lead to a substantial increase in th ntent of the pellets. The increase in final moisture content with increa idity is also displayed in Figure 2.12. t 50% relative humidity, the final moisture contents of the pellets at eratures of 15, 25 and 35?C were 7.84, 7.92 and 7.94%, respectively. However, at inal moisture content was much lower than that obtained at other temp ). It is postulated that the elevated temperature (45?C) is causing the isture to evaporate near the surface of the pellets before it can be absorbed. 6 15 20 25 7 8 9 10 11 12 0 5 10 30 35 40 45 T me (h) Mo i s t u re co n t en t ( % , w . b . ) 50% 65% 80% Predicted i chgrass pellets exposed to air at 15?C and relative Figure 2.12 ? Moisture change in swit humidity of 50, 65, and 80%. Initial moisture content was 7.19% (w.b.). 6 7 9 10 11 12 25 me (h) o e n te n t (% . , w b . ) 55 8 0 5 10 15 20 M i s tu r C o 15?C 25?C 35?C 45?C Predicted Ti Figure 2.13- Moisture change in switchgrass pellets exposed to air at 50% relative m t resistance is concentrated in a layer at the surface of the material. humidity at temperatures of 15, 25, 35, and 45?C. Initial moisture content was 7.19% (w.b.) The non-linear estimation procedure (NLIN) in SAS 9.1 Statistical package (2002) was used to fit an exponential model (Equation 2.21) to the moisture sorption data. The estimated values for moisture sorption rate constant (k) and equilibrium oisture content (M q ) are given in Table 2.3. Equation 2.21 is one of the simplest models that has been used to describe moisture sorption in biological materials (Fasina and Sokhansanj, 1992). According to Chhinnan, (1984) this model assumes negligible internal resistance which means there is no resistance to moisture movement from within he material to the surface. It only considers the surface resistance implying all the )exp( kt MM MM qi q ?= ? ? (2.1) Where, t = time (min) M = instantaneous moisture content (%, w.b.) M i = initial moisture content (%, w.b) Table 2.3- Estimated values of moisture sorption rate constant (k) and equilibrium moisture content (M q ) obtained from non-linear regression analysis using Equation 2.21 for switchgrass pellets RH (%) Temperature (?C) k (h -1 ) M q (%, w.b.) 15 0.33 7.84 50 25 0.35 7.92 35 0.52 7.94 45 0.42 7.44 15 0.28 8.26 65 25 0.23 8.83 35 0.24 9.02 45 0.38 8.70 15 0.13 11.09 80 25 0.18 11.18 35 0.20 10.90 The values of the coefficient of determination, R 2 , were between 0.893 and 0.994 and values of standard error were between 0.0025 and 0.00574, which showed that Equation 2.21 adequately predicted the experimental data. Further statistical analysi indicated the estimated values of k and M s d (P < 0.05) by lative humidity and not by temperature (Table 2.2). The k value was higher at lower relative humidity and higher with incr dependency of k and M q on relative h midity with other biological materials (Zink, 1935; q were significantly affecte re easing temperature. Other researchers reported the u 56 57 ar et al., 2004; McMullen et al., 2005). At higher relative longer. The pellets gained up to 4 . Data for other temperature and relative umidity combinations can be found in Appendix A, Figures A.1-A.4. M The EMC-ERH curves were sigmoidal (Type II) in shape which is typical of equilibrium moisture content data for biological m (Lab as et al., 2005). Gene , the relative hu ity at given moisture content increased with increasing te perature until RH 72 and M e = 11%. After this p he effect of temperature was insignificant (Figure 2.14). According to van den Berg and Bruin (1984), the isotherm curve can be divided into three regions. The first region has a range of relative humidity from 0 to 0.2 with an enthalpy porization larger than pure water. In this region the first wa olecules are s at the act r sites. Consequently, the molecules are tightly bound and behave as part of the solid. In the second region with a relative humidity range of 0.23 to 0.73 the water is less tightly bound and the enthalpy of vaporization is closer to that of pure water. The water molecules sorb near to or on top of the first molecules and penetrate into newly formed holes of the swollen structure. Depending on the solids present this water enhances chemical and biochemical reactions. In the third region, with relative humidity range of 0.74-1.0, there is free water, which is mechanically trapped in the void spaces of the solid and has properties similar to those of bulk water. It is in this third region that the overlapping of isotherms is observed in this study. The isotherms show temperature has Dexter et al., 1947; Dadg humidity values, the time for pellets to equilibrate was percentage points of moisture at high humidity h oisture Sorption Isotherms a lsteria uza 1 Erb984; rally mid m = 0. oint, t of va much ter m orbed ive pola 58 and onstant moisture content is greater at low to intermediate relative humidity (Labuza, 1984). As seen in igure 2.14, relative humidity increases as temperature increases for a given equilibrium moisture content. This can make the pellets susceptible to microbial growth. little or no effect on equilibrium moisture content at high relative humidity. Temple van Boxtel (1999) reported similar results with black tea leaves. It can be noted that materials that follow a Type II isotherm hold less water at higher temperatures. In general the effect of temperature on increasing the relative humidity at c F 0 5 15 0 0.2 0.4 0.6 0.8 1 10 20 Relative Humidity ilb m is d E q u r i u M o tu r e (% , . b . 6?C 20?C 35?C 50?C The EMC and ERH data were analyzed using the following four equations (Fasina and Sokhansanj, 1993): The Halsey equation: Figure 2.14 ? Moisture sorption isotherms for switchgrass pellets at 6, 20, 35, and 50?C ( )[ ] c? MbtaRH +?= expexp (2.22) 59 The Modified Henderson equation: 1? ? ? ? ? ? ? ? ? ? ? + c M bta 1 ?? + ?? =RH (2.23) The Chung-Pfost equation: () ? ? ? ? + ? bt a The Oswin equation: ?? ?= cMRH expexp (2.24) ( )[ ] c MbtaRH +??= exp1 (2.25) here, RH is the relative humidity in decimal, M is the equilibrium moisture content on The standard error of estimate (SEE) is the conditional standard deviation of the ependent variable and is defined as follows: W a dry basis (which is a typical way or reporting isotherm data) and a, b, and c are model coefficients. The coefficients of the equations were estimated using the non-linear regression procedure (NLIN) in SAS package 9.1. The goodness of fit for each model was quantified using the coefficient of determination (R 2 ), the standard error of the estimate (SEE) and the mean relative deviation (MRD). Several authors have reported the benefit of using multiple statistical parameters to select the model that gives the best fit to experimental data (Fasina and Sokhansanj, 1993; Barker, 1996; Soysal and Oztekin, 2001; Erbas et al., 2005, Durakova et al., 2005; Pagano and Masheroni, 2005). d () df MM SEE i ee? = ? = 1 (2.6) Where, M e is the measure value, M ? is the estimated value, m is the number of data points and ?df? is the degrees of freedom for the model. The mean relative deviation m 2 ? e 60 (MRD) is an absolute value used because it re of the mean divergence of the estimated data from the m gives a clear pictu easured data and is defined as follows: ? ? ? ? ? ? = m ee MM m MRD ? 1 (2.7) In general, when evaluating the goodness of fit of a model, low values of R 2 and high values of SEE and MRD mean the model is unable to accurately describe the variation in experimental data. The model coef or a temperature range = ? ? ? ? ? ? i e M 1 ficients and statistics of fitting f of 6-50?C are listed in Table 2.4. It can be observed that the Halsey equation gave the highest values for MRD and SEE and the lowest R 2 value suggesting that it is not the most appropriate model to describe the experimental data (Figure 2.15 (a)). The Chung- Pfost model gave the lowest MRD, and SEE and the highest R 2 value leading to the appearance of the most suitable model (Figure 2.15 (c)), which follows the conclusion by Soysal and Oztekin (2001) that the Chung-Pfost equation is suitable for cereal and other field crops. However, is has been noted that isotherm equations that gave values of MRD<0.05 have been considered to be a good fit (Pagano and Mascheroni, 2005). All of the equations except the Halsey equation had MRD values less than 0.05. 61 standard error of estimate (SEE) and coefficient of determination (R ) for temperature Statistics of fitting Table 2.4 ? Model coefficients and values for the mean relative deviation (MRD) 2 range of 6-50 ?C Coefficients Equation a c MRD SEE Rb 2 Halsey 3.6422 -0.0040 1.8172 0.1279 0.1737 0.9789 Stdev ? 0.1785 ?0.00176 ? 0.0760 M. Henderson 3.70E-05 34724.0 1.7618 0.0484 0.0744 0.9882 Stdev ? 0.000017 ? 163.5 ? 0.0572 Chung-Pfost 1083.9 218.60 0.2021 0.0266 0.0553 0.9935 Stdev ? 256.8 ? 57.62 ?0.00476 Oswin 9.3788 -0.0169 2.6285 0.0387 0.0569 0.9931 Stdev ? 0.1570 ? 0.00473 ? 0.0673 62 0 0 0.2 5 10 15 0.8 R E q u ili b r iu m M , d 20 . b . ) 0.4 0.6 1 elative Humidity o is t u re ( % 0 10 .2 0 Rel e Humidit E br i um M o e ( 5 15 20 % , d. b. ) 00 0.4 .60.81 ativ y qul i i s t ur (b) 0 5 10 0 0.2 0.4 0.6 0.8 1 Relative Humidity E qui l i br i u r e ( % , d. b . ) 15 20 m M o i s t u (c) 0 5 10 1 0 0.2 0.4 0.6 0.8 1 Relative Humidity E qui l i br i um M o i s t u r e ( % , d. b . ) 5 20 (d) Figure 2.15 ? EMC/ERH data fitted to four isotherm models: (a) Halsey equation (b) Modified Henderson equation (c) Chung-Pfost equation (d) Oswin equation. (?, experimental; ???, model) 63 It can be concluded from affects physical properties of pellets ma isture content increased the diameter of the pell Bulk and particle densities decreased by 24 a the pellets increased. A maximum durability medium durability in the moistu a result of increasing moisture. The force 20.60 to 30.21N. There was a maximum moisture content. This moisture level could isture content for switchgrass pellets Results from compositional analysis reveal compo er, atistically there were no differences in carbohydrate levels, but there were significant ifferences in ash, acid soluble lignin and acid insoluble lignin at a 95% confidence terval. There is no scientific explanation for the increase in proportion of components f the sample due to pelleting Relative humidity had a significant effect on the moisture uptake rate constant (k) nd equilibrium moisture content (M q ). At lower relative humidity, k was higher. The ellets also gained 4% in moisture at high relative humidity. The moisture sorption otherms showed independence to temperature beyond the relative humidity of 0.72. For CONCLUSION this study that moisture content significantly nufactured from switchgrass. Increasing mo ets by 8% and decreased the length by 17%. nd 16% respectively as moisture content of rating of 95.91% was obtained when the pellets were at a moisture content of 8.62% (wet basis). Pellets also displayed high to re range evaluated. Durability and hardness decreased as required to rupture the pellets ranged from or minimum value of all properties at 8.62% possibly be considered an optimum mo upon further investigation. ed that proportion values of individual nents for the pelleted sample were higher than those of ground sample. Howev st d in o a p is 64 a temperature range between 6 and 50 st equation was the most approp iate. ?C the Chung-Pfo riate fit to experimental data where the Halsey equation was the least appropr 65 Aarseth, K. A., E. Prestlokken, E. (2003). Mechanical properties of feed pellets: Weibull dapa, P. K., Schoenau, G. J., Tabil, L. G., Sokhansanj, S., Crerar B.J. (2003). Pelleting of fractionated alfalfa products. ASABE Paper No. 036069: pages 1-11, American ociety of Agricultural Engineers Annual international Meeting, July 27-30, Las Vegas, Nevada, USA. ouri, F., Gilbert, T. I., Dale, B. E. (2005). Pretreatment of itchgrass by ammonia fiber explosion (AFEX), Applied Biochemistry and Biotechnology, 121-124, 1133-1142. ASABE (2002). Cubes, pellets, and crumbles--definitions and methods for determining ensity, durability, ASABE Standard, S269.4, St. Joseph, MI. SABE (2003). Method of determining and expressing fineness of feed materials by sieving, ASABE Standard, S319.3, St. Joseph, MI. anian, D. (2001). Physical properties of raw cashew nut, Journal of gricultural Engineering Research, 78, 291-297. arger, P.C. Biomass Transport Systems, In: Encyclopedia of Agriculture, Food and Biological Engineering, Edited by Heldman, D., New York: Marcel Dekker Publisher; 003, 94-98. uilibrium moisture content of cotton plant components, Journal f Agriculture Engineering Research, 63, 353-364. hhinnan, M. S., (1984), Evaluation of selected mathematical models for describing thin rying of in-shell pecans, Transactions of the ASABE, 27, 610-615. nction of temperature and relative humidity during storage, An ASABE/CSAE Meeting American Society of Agricultural Engineers/Canadien ociety of Agricultural engineers, Annual International Meeting, August 1-4, Ottawa, (1993). Physical properties of soybean, Journal of gricultural Engineering Research, 56, 89-98. exter, S. T., Sheldon, W. H., Waldron, D. I. (1947). Equilibrium Moisture content of Agricultural Engineering, 28, 295-296. REFERENCES analysis, Biosystems Engineering. 84, 349-361. A S Alizadeh, H., Teym sw d A Balasubram A B 2 Barker, G. L. (1996). Eq o C layer d Dadgar, S., Tabil, L. G., Crerar, B. (2004). Moisture Adsorption of files peas as a fu Presentation, Paper NO. 046089, S Ontario, Canada. Deshpande, S. D., Bal, S., Ojha, T. P. A D alfalfa hay, 66 Durkova, A. G., N.D. Menkov N. D. orption characteristics of rbas, M., Ertugay, M. F., Certel, M. (2005). Moisture absorption behavior of semolina the dilute sulfuric a id pretreatment of corn stover, poplar and itchgrass, Bioresource Technology, 59, 129-136. alfa cubes , 35, 1615-1619. ,. 56, 51-63. lfa eting of ASABE, Kansas City, MO, ne 19-22. andling characteristics of alfalfa ellets, Powder Handling and Processing, 8, 361-365. ., Gordon and Breach Science, Amsterdam, 1989, 860 ? 891. ngineers Annual International Meeting July 27-30, Las Vegas, Nevada, USA. (2005). Moisture s chickpea flour, Journal of Food Engineering, 68, 535-539. E and farina, Journal of Food Engineering, 96, 191-198. Estghlalian, A., Hashimoto, A. G., Fenske, J. J., Penner, M. H. (1997). Modeling and optimization of sw Fasina, O.O., Sokhansanj, S. (1992). Hygroscopic moisture absorption by alf and pellets, Transactions of the ASABE Fasina, O.O., Sokhansanj, S. (1993) Equilibrium moisture relations and heat of sorption of alfalfa pellets, Journal of Agricultural Engineering Research Fasina, O.O., Sokhansanj, S., Irudayaraj, J. (1994). Thermo-physical properties of alfa pellets, Paper No. 946022, Annual International Me Ju Fasina, O.O., Sokhansanj, S. (1996). Storage and h p Jenkins, B.M., Physical Properties of Biomass, In Biomass Handbook, Editors. Kitani, O., Hall, C.W Joshi, D. C., Das, S. K., Mukherjee, R. K. (1993). Physical properties of pumpkins seeds, Journal Agricultural Engineering Research, 54, 219-229. Khoshtaghaza, M. H., Sokhansanj, S., Gossen, B. D. (1999). Quality of alfalfa cubes during shipping and storage, Applied Engineering in Agriculture. 15, 671-676. Labuza, T. P. Moisture sorption: aspects of isotherm measurement and use, The American Association of Cereal Chemists, ?1984. Lehtikangas, P. (2001). Quality properties of pelletised sawdust, logging residues and bark, Biomass and Bioenergy, 20, 351-360. Liu, C., Wyman, C.E. (2005). Partial flow of compressed hot water through corn stover to enhance hemicellulose sugar recovery and enzymatic digestibility of cellulose, ioresource Technology, 96, 1978-1985. B Mani, S., Tabil, L. G., Sokhansanj, S., Roberge, M. (2003a). Mechanical properties of orn stover grind, ASABE paper No. 0.36090, American Society of Agricultural c E 67 ., hgrass as a bioenergy crop, : Perspectives on new crops and new uses, Edited by Janick J., Alexandria, VA: ASHS cLaughlin, S.B., Kszos, L.A. (2005). Development of switchgrass (Panicum virgatum) (2005). Storage and handling re, Ed). New York: Gordon and s and other rops, Transactions of the ASABE, 45, 165-170. l properties of green gram, hington tistics: SAS Institute Inc., Cary NC. McLaughlin, S.B., Bouton, J., Bransby, D., Conger, B., Ocumpaugh, W., Parrish, D Taliaferro, C., Vogel, K., Wullschleger, S. Developing switc In Press; 1999, p. 282?299. M as a bioenergy feedstock in the United States, Biomass and Bioenergy, 24, 515-535. cMullen, J., Fasina, O.O., Wood, C. W., Feng, Y. M characteristics of pellets from poultry litter, Applied Engineering in Agriculture, 21, 645- 651. McNeill, S.G., Thompson, S.A., Montross, M.D. (2004). Effect of moisture content and broken kernels on the bulk density and packing of corn, Applied Engineering in griculture, 20, 475-480. A Missaoui, A., Fasoula, V., Bouton, J. (2005). The effect of low plant density on response to selection for biomass production in switchgrass, Euphytica, 142, 1-12. Mohsenin, N. N. (1986). Physical Properties of Plant and Animal Materials: Structu Physical Characteristics and Mechanical Properties (2 nd Breach Publishers. Nelson, S. O. (2002). Dimensional and density data for seeds of cereal grain c imkar, P.M., Chattopadhyay, P. K. (2001). Some physicaN Journal of Agricultural Engineering Research, 80, 183-189. REL (2004). Laboratory Analytical Procedures - Determination of Structural N Carbohydrates and Lignin in Biomass, United States Department of Energy, Was D.C. NREL (2005). Laboratory Analytical Procedure - Determination of ash in biomass, United States Department of Energy, Washington D.C. Pagano, A. M., Mascheroni, R. H. (2005). Sorption isotherms for amaranth grains, Journal of Food Engineering, 67, 441-450. AS (2002-2003). SAS Users? Guide: StaS Sokhansanj, S., Turhollow, A. F. (2004). Biomass densification ?cubing operations and costs for corn stover, Applied Engineering in Agriculture, 20, 495-499. 68 abil, L. G., Sokhansanj, S. (1996). Process conditions affecting the physical quality of abil, L. G., Sokhansanj, S. (1997). Bulk properties of alfalfa grind in relation to its icultural Engineering Research, 74, 83-89. iselogel, A. E., Agblevor, F. A., Johnson, D. K., Deutch, S., Fennell, J. A., Sanderson, oodcock, C.R. and Mason, J.S., Bulk solids handling: an introduction to the practice e hays, Agricultural Engineering, 16, Soysal, Y., Oztekin, S. (2001). Comparison of seven equilibrium moisture content equations for some medicinal and aromatic plants, Journal of Agriculture Engineering Research. 78, 57-63. T alfalfa pellets, Applied Engineering in Agriculture, 12, 345-350. T compaction characteristics, Applied Engineering in Agriculture, 13, 499-505. Temple, S. J., van Boxtel, A. J. B. (1999). Equilibrium moisture content of tea, Journal f Agro Thomas, M., van der Poel, A. F. B. (1996). Physical quality of pelleted animal feed 1. Criteria for pellet quality, Animal Feed Science Technology, 61, 89-112. van den Berg, S. Bruin Water activity and its estimation in food system: theoretical aspects, Editors: L. B. Rockand, G. F. Stewart. Academic Press Inc. New York. 1981. W M. A. (1996). Compositional changes during storage of large round switchgrass bales, Bioresource Technology, 56, 103-109. W and technology. Leonard Hill, New York, 1987. um moistures of somZink, F. J. (1935). Equilibri 451-452. 69 ITCHGRASS Biomass materials are renewable sources of energy. They are derived from living her biological d oncerns about environmental pollution from fossil fuel utilization and to secure our ch as biomass. Similar to other biological materials, low bulk density is a characteristic nsportation very expensive (Mani et al., ulk density thereby creating simpler handling and reduced transportation and storage osts (Smith et al., 1977; Mani et al., 2003b; Mani et al, 2005). Compaction (or densification) involves the application of mechanical pressure to reduce the volume of biological material. Extrusion and briquetting are the most common commercially employed methods. Extrusion processes use pistons, screws, or rollers to force the material through one or more constrictive dies thus resulting in the production of pellets, cubes, or wafers (Bruhn, 1989). Extrusion processes consume high amounts of energy due to the excessive frictional resistance. This resistance causes the system to require a high power input which generates elevated temperatures in the CHAPTER 3 ? COMPACTION BEHAVIOR OF SW INTRODUCTION plants, animal manures, waste products from processing industries and ot sources (Mani et al., 2003a). Due to the increasing cost of energy from fossil fuel an c nation?s food supply, there is an urgent need to obtain energy from renewable resources su that makes handling difficult, and storage and tra 2005). Compaction is the most common method that has been employed to increase the b c 70 pelleted materials (Faborode, 1990). Materials that are compacted into pellets are usually denser than other compacted forms (i.e. briquettes, cubes and wafers). Pelleted materials are therefore preferred when there will be significant handling and transportation of the pacted material. In order to manufacture high quality pellets, knowledge of ompaction behavior is very important. Converting biomass to pellets or briquettes is greatly influenced by processing par ture content, temperature and die size e size of 6.2 mm is needed to pelle up to com c ameters such as mois . Compaction studies can aid in optimizing processing parameters that will result in the production of high quality pellets at pilot and industrial scales. Smith et al (1977) reported that an increase in temperature led to a greater degre of compaction, and recovery length of the straw briquettes was less when applied temperatures were in the range of 90 to 140?C. Tabil and Sokhansanj (1996) reported that it was important for alfalfa grinds to reach above 90?C after passing the conditioning chamber to ensure high quality pellets. The adhesive properties of thermally softened lignin are thought to contribute significantly to the strength characteristics of briquettes made from lignocellulosic materials (Grenada et al., 2002). Tabil and Sokhansanj (1996) found that an optimum die t alfalfa grind with a moisture content of 7.5-9.0% moisture. They also found that a die size of 7.9 mm could be used to pellet alfalfa grinds with moisture content of 12.0%. This is because alfalfa grinds are likely to be gummy when hot and moist. The smaller diameter dies offer more resistance to the grind particles due to a lower surface area, hence an increase in density. In addition, during the pelleting of alfalfa, steam is added to aid the release and to activate natural binders and lubricants thereby increasing 71 Most nd hn, 1989; Mani et al., 200 . The e MATERIALS AND METHODS The switchgrass used in this study was obtained from the E.V. Smith Experiment Station (Auburn University) in Tallassee, Alabama. Before experimentation, the switchgrass was ground in a hammer mill (New Holland Grinder, Model 358, New Holland, PA) that was fitted with a screen size of 3.18 mm. The compression test was conducted using a single pellet apparatus that consists of a cylindrical jacketed die pellet durability and reducing energy consumption during pelleting (Tabil and Sokhansanj, 1996; Samson et al., 2000). The energy required for compaction of biomass depends highly on pressure, moisture content, physical properties of the material and method of compaction. densification processes encompass two main components, the energy of compression, a the energy of extrusion. The compression energy results in bulk reduction and the extrusion energy is required to overcome the friction between the material and the die, and may account for roughly 60% of the total densification energy (Bru 4a). There have been reported efforts to pellet switchgrass but no parametric studies to investigate the die size, moisture content, or temperature during pelleting aim of this study was to quantify the effects of moisture content, temperature, and die size on the degree of compaction that can be obtained from switchgrass in a single pellet apparatus and the specific energy required to achieve this degree of compaction. Limited studies were also carried out on the effect of compaction force and particle size on degre of compaction. 72 e 4.8, 6.4, and 7.9 mm), moisture content (10.4, ssion ng ie used to compress (at a speed of 1 mm/s) the sample in the die to a force ad of 3924 N. Once this force was attained, the software provided by manufacturer of the Texture Analyzer was programmed to hold the sample at this force (stress relaxation) r a period of 1 minute (Figure 3.2). After this duration of time, the pellet was removed from the die. The weight (using a balance accurate to 0.001 g ,Model AR3130 Ohaus Corp, Pinebrook, NJ) and the dimensions (length and diameter using a digital caliper accurate to 0.01 mm , Model CD-56C, Mitutoyo Corp, Kawasaki, Japan) of the pellet were then obtained. During sample compression, force-deformation-time data from the Texture Analyzer were automatically displayed and recorded by the software. (diameter s of 4.8, 6.4, and 7.9 mm and length of 50 mm). Parameters that wer investigated in the study include die size ( 13.2, 16.2 and 20.0%, wet basis.), and temperature (60, 75, and 90?C). The moisture content of the ground switchgrass was adjusted to the desired moisture levels by putting samples in an environmental chamber (Model AA-5460A, Espec Corp., Hudsonville, MI) set to 30?C and 90% relative humidity. The compre apparatus was composed of a plunger and die assembly attached to a texture analyzer (Model TA-HD, Stable Micro Systems, Surrey, UK) (Figure 3.1). Water from a water circulator (Model 9512, Polyscience, Niles, IL) was used to heat the samples by flowi through the annular space of the jacketed die. To conduct a test, a weighed amount of ground switchgrass was placed in the d and heated for 20 minutes to the desired temperature before compacting. The time duration of 20 minutes was verified by monitoring and recording the temperature reading from a thermocouple that was placed in the sample during the heating process. The plunger was then lo fo 73 es of e compression of each pe f The density of each pellet was computed from Equation 3.1 using the valu the measured mass and volume of the pellet obtained from the die. The area under the force-deformation curve was taken as the specific energy required for th llet (Figure 3.3). Limited studies were also carried out on (a) compaction of switchgrass using the different die sizes (4.8, 6.4, and 7.9 mm) with the same amount o pressure (95 MPa) and (b) on effect of particle size on compaction using hammer mill screen sizes of 0.79, 1.58, and 3.18 mm. pellet pellet V m =? (3.1) pellet Figure 3.1- Single pellet compaction test set up 0 1500 4000 80 120 160 Time (s) F 500 1000 2000 2500 3000 3500 0 40 o rce ( N ) Compression Relaxation Figure 3.2 ? Force-time ade in single pellet apparatus (conditions: 4.8 mm, 60?C, 10.4%, w.b.) curve for switchgrass m 0 500 1000 1500 2000 2500 3000 3500 4000 0 0.08 0.1 F o rce ( N ) 0.02 0.04 0.06 Distance (m) 74 Figure 3.3 ? Force-distan uired to compress the sw ce curve used to calculate energy req itchgrass (conditions:4.8mm, 60?C, 10.4% w.b.) 75 sam biom tem bioma de 1990; Tabil and Sokhansanj, 1996; Samson et al., 2000). RESULTS AND DISCUSSION Figure 3.4 shows the temperature data obtained from the thermocouple placed in the sample during the heating process. The figure shows that it took less than 5 minutes for the samples to reach each desired temperature. Increasing the temperature of the ple is crucial to the compaction process because switchgrass is a lignocellulosic ass. Several studies on similar lignocellulosic biomass (e.g. alfalfa) have shown that the fibrous component of this biomass type has to be melted (achieved by increasing perature) before significant compaction (or increase in density) and stability of the ss can be achieved eeler, 1984; Faboro (Smith et al., 1977; O?Dogherty and Wh 0 20 40 60 80 0 5 10 15 20 25 Time (min) T e m p era t u re ( ? C 100 ) 60?C 75?C 90?C Figu .b.) Density The effect of die size, temperature and moisture content on the density of pellets obtained from the single-die apparatus is illu rated in Figure 3.5. Statistical analysis revealed that all the variables (die size, temp oisture content) significantly re 3.4 ? Temperature profiles of switchgrass sample at 13.2% moisture content (w during conditioning st erature and m 76 affected (P < 0.05) pellet den icant interaction between variabl ce perature on the fibrous com Sams within th er 60 terial. It ide a latively high m nd hay before compressing allows greater package density and durability to be achieved. Sokhansanj (1996) found that it was important for alfalfa grinds to reach a temperature above 90?C during conditioning to ensure pellet quality and efficiency. It was shown (in the same study) that the conditioning temperature had a positive sity. There was however, no signif es. As expected, the density of the switchgrass pellets decreased as die size increased from 4.8 to 7.9 mm diameter regardless of the temperature and moisture content combination (Figure 3.5). This was primarily due to (a) increase in surface area (hen lower compaction pressure) as die size increased, and (b) greater folding and interlocking of the particles as die size decreased (Butler and McColly, 1959; O?Dogherty and Wheeler, 1984). Density of the pellets increased within the temperature range of 60 to 90?C. It is suspected that this is due to the documented effects of tem ponents of lignocellulosic biomass such as switchgrass. This is in agreement with on and Duxbury?s (2000) findings which stated that between 75-85?C, the fiber e cellulosic biomass begins to flow from the fiber cell wall and binds with oth fibers during compression. Other researchers showed that heating a material between and 70?C led to a more stable product than what was possible with unheated ma was also determined that at higher temperatures, lower pressure was needed to prov given degree of compaction for lucerne and Bermuda grass and that grass with re oisture content could be stably compacted at elevated temperatures (Smith et al., 1977). O?Dogherty and Wheeler (1984) reported that increasing the temperature of straw a Tabil and 77 ith durability (see results in Chapter 2). the over briquettes (Mani et al., 2004b). The decrease in density is thought way y correlation with the durability of the manufactured alfalfa pellets where density is highly correlated w As moisture content increased, the density of the pellets decreased. For example when the moisture content of switchgrass samples increased from 10.4% to 20%, density of pellets (made at 90?C and die size of 4.8 mm) decreased from 1215 kg/m 3 to 955.21 kg/m 3 , respectively (Figure 3.4). Similar results were reported by for wheat straw briquettes (Smith et al., 1977), straw wafers (O?Dogherty and Wheeler, 1984) and corn st to be due to the inhibitory effect of the water molecules on particulate bonding during the compaction process. It is postulated that moisture present which cannot escape, by of extrusion through vent holes and/or clearance between the compression chambers ma limit the maximum dry matter density attainable, and may interfere with the performance of natural bonding agents (Bruhn, 1989). High moisture can also cause axial expansion of the pellets, thereby reducing density (Mani et al., 2005). 78 a) 600 800 60 75 90 Temperature (?C) D e n 1000 1200 t y ( g/ m 3 1400 s i k ) 4.8mm 6.4mm 7.9mm b) 600 1400 60 75 90 Temperature (?C) D t g/ 800 1000 1200 e n s i y ( k m 3 ) 4.8mm 6.4mm 7.9mm Figure 3.5 ? Density of switchgrass pellets affected by temperature and die size at constant moisture contents of a) 10.4% and b) 20.0% In this study, the maximum density attained when switchgrass was compacted to a force of 3924 N in a single pellet apparatus was 1214 kg/m 3 at 10.4% moisture content and 90?C. Mani et al (2004b) obtained a maximum corn stover briquette density of 950 kg/m 3 in the moisture range of 5-10% (wet basis). Other researchers found the optimal 79 moisture content for lucerne and Bermuda grass to be between 16 and 23% (wet basis) (Sm com when biological materials are compacted. Butler and McColly (1959) confirmed that at a W ith et al., 1977). It is therefore obvious that the optimum moisture content required to produce high density compacts varies with the type of feedstock. Effect of Pressure and Grind Size Based on the studies that were carried out, it was determined that application of the same amount of pressure (95 MPa) to samples in different die sizes did not result in pacts of the same density. The density of the compacts (Figure 3.6) decreased with increase in die size thereby confirming that die size is as important as applied pressure given pressure the density of hay pellets were greater for smaller diameter chambers. Other studies reported the pressure required to form hay wafers (O?Dogherty and heeler, 1984) and palm fiber briquettes (Husain et al., 2002) of a given density increased exponentially with die diameter. It is stipulated that as the die becomes larger there is more relaxation of the particles resulting in less folding and interlocking of particles. 80 400 Die Size D 600 800 1200 4.8 mm 6.4 mm 7.9 mm e n k g/ 1000 m 3 ) s i t y ( Figure 3.6 ? Density of switchgrass pellets with constant applied pressure of 95 MPa, 10.4% moisture Results from the study on effect of pellet density on grind size (Table 3.1) show that the density of the compacts increased as the particle size decreased (Figure 3.7 Further statistical analysis (at 95% significance level) indicated that the density va from sample A (1274 kg/m ed ). lues e es of samples A and B. The samples used in this part of the study were btained by using the hammer mill (referenced in Materials and Methods) to grind the samples through screen sizes of 0.79, 1.59, and 3.18 mm. The geometric mean diameter (d gw ) and standard deviation (S gw ) of the samples were obtained according to ASABE Standard S319.3 (2003) (Table 3.1). 3 ) were not significantly different from that of sample B (1272 kg/m 3 ). Density values from sample C (1214 kg/m 3 ) were significantly lower that th density valu o Table 3.1 ? Geometric mean diameter and standard deviation of the samples from the three hammer mill screen sizes Sample Screen size (mm) d gw (mm) S gw (mm) A 0.79 0.191 0.140 B 1.59 0.231 0.127 C 3.18 0.860 0.370 800 900 0.79 1.59 3.18 Screen Size (mm) D e 1000 1200 1300 n s i t y ( k 3 ) 1100 g/ m AB C Figure 3.7 ? Density of switchgrass pellets as particle size increase The size of feed grind influences the final quality of densified masses. Similar results on the effect of grind size on pellet quality have been reported by other researchers (Lehtikangas 2001; Mani et al., 2003a; Adapa et al., 2004). Fine or medium ground feed constituents are desirable for pelleting because more surface area is available for moisture addition during steam conditioning (Tabil and Sokhansanj, 1996; Adapa et al., 2004). During compaction, smaller (fine) particles rearrange and fill in the void space of larger (coarse) particles producing denser and more durable compacts. Coarsely ground materials tend to give less durable pellets because they may create natural fissures (conditions: 90?C, 4.8 mm) 81 82 in is essential, but if r material is prese quality a efficiency will suffer. It has been reported th ng the b material ller size results in higher density briquett ith better d ity, lowe abs rates, and an increase in binding pro ties o the f ck (Sam l., from m comp switchgra energy whi nergy required for compression of sw nge of 19.33 - 88.81 MJ/t epend energy the pellet, which are susceptible to breakage. A proportion of fine and medium grind too little coa se nt, nd at grindi iomass to sma particle es w urabil r water orption per f eedsto son et a 2000; Jannasch et al., 2001b; Lehtikangas, 2001). Mani et al (2003a) concluded that low quality alfalfa grinds a 2.4 mm hammer mill screen had higher cohesion than the grind from a 3.2 m screen. Also, particles with sizes below 0.400 mm are considered fine and highly ressible (Mani et al., 2003a). Specific Energy Statistical testing on the specific energy (Figure 3.7) required to compact ss indicates that moisture and die size had significant effect on compaction le temperature was not significantly related to energy. The specific e itchgrass pelle s was in the rat d ing on test parameters. Tabil and Sokhansanj (1996) reported a specific energy required for alfalfa pellets of 106.92 MJ/t. Mani et al (2004b) reported specific for compression of corn stover briquettes in the range of 7.31-16.08 MJ/t. O?Dogherty and Wheeler (1984) reported specific energy required for straw wafers in the range of 5.5 - 21.3 MJ/t. The method of densification along with test parameters dictates the amount of energy required. Pellets are more compact than briquettes and require more energy to achieve the maximum density attainable. 83 al., ontents of <10% (wet basis) for lignocellulosic material. Anything above 10% can lead to erratic pelleting, less dense compacts, and more required energy. Similar results were reported with alfalfa hay pellets (Butler and McCol les is or Specific energy decreased by 66% as die size increased from 4.8 to 7.9 mm (Figure 3.8). Less energy was required to compress the switchgrass as the die size increased from 4.8 mm with a range of 88.81 ? 44.34 MJ/t to 7.9 mm with a range of 29.26 ? 19.33 MJ/t. Specific energy required to compact samples in the 4.8 mm diameter die increased with moisture content. It is suspected that less space for relaxation gave rise to an increase of energy required to overcome inhibitory forces stemming from moisture gain. Smaller dies have been reported (O?Dogherty and Wheeler, 1984; Samson et 2000) to be able to handle moisture c ly, 1959; Bellinger and McColly, 1961). Unlike the result from the 4.8 mm diameter, specific energy required for samp densified in 6.4 and 7.9 mm diameter dies decreased as moisture content increased. Th may be due to the increased surface area of the die, which leaves more available space f relaxation of the particles and less resistance to compression (Figure 3.8). For pellets made at 16.2% moisture content, results were very sporadic with no obvious trend. 84 (a) 0 60 20 40 80 100 60 75 90 4.8mm 4.8mm 4.8mm Temperature (?C) e n e (M S p c i fi c E r g y J / t) 10.4% 13.2% 20.0 (b) 0 20 40 80 100 60 7 60 5 90 6.4mm 6.4mm 6.4mm Temperature (?C) c y t) S p e c i fi E n e r g (M J / 10.4% 13.2% 20.0% (c) 0 20 60 60 75 90 7.9mm 7.9mm 7.9mm 40 80 100 Temperature (?C) S p e c i fi c t) En e r gy (M J/ 10.4% 13.2% 20.0% Figure 3.8 ? Effect of temperature, moisture content and die size on specific energy required to make switchgrass pellets in a single pellet apparatus at constant die sizes of (a) 4.8 mm (b) 6.4mm and (c) 7.9 mm (mean values are reported) 85 parame ressing comp denser com amount same increase in d 7.9 mm relax and resistan ntent comp CONCLUSION A relationship between the density and specific energy required and processing ters (die size, moisture content, and temperature) was established by comp ground switchgrass using a single pellet apparatus. The density of the switchgrass acts decreased with increasing die size resulting from the increase in surface area. There was less relaxation of the material as the die size became smaller, and therefore, a pact was formed in the smaller (4.8 mm) die. Application of the same of pressure (95 MPa) to samples in different die sizes does not result in compacts of the density. In addition the density of the compacts increased as particle size deceased. Specific energy required to compress the ground switchgrass decreased with ie diameter. For die diameter of 4.8 mm there was an increase in energy required with increasing moisture content. The opposite occurred with 6.4 and die sizes where the specific energy required decreased as moisture content increased. This may be due to the increase in surface area which allows particles to ce to compression is lessened. Because results from 16.2% moisture co data were sporadic it is concluded that a moisture level of 16.2% should not be used for action of switchgrass. 86 dapa, P. K., Tabil, L. G., Schoenau, G. J., Sokhansanj, S. (2004). Pelleting charact Engineering in Agriculture, 20, 813-820. ASABE (2003). Method of determining and expressing fineness of feed materials by Agriculture Engineering, 42, 244-247. Bruhn, H.D., Mechanical treatment, In Biomass and Handbook, ed. Kitani,O. and Hall., Engineering, 40, 442-446. Faborode, M. O. (1990). Analysis of extrusion compaction of fibrous agricultural lignocellulosic briquettes, die design, and products study, Renewable Energy, 27, 561- the processing of palm nuts to palm oil, Biomass and Bioenergy, 22, 505-509. Jannasch, R., Quan, Y., Samson, R. (2001b). A process and energy analysis of REFERENCES A eristics of fractionated sun-cured and dehydrated alfalfa grinds, Applied sieving, ASABE Standard, S319.3, St. Joseph, MI. Bellinger, P. L., McColly, H. F., (1961), Energy requirements for forming hay pellets, C.W., Gordon and Breach Science, Amsterdam, 1989, 452-459. Butler, J. L., McColly H. F., (1959), Factors affecting the pelleting of hay, Agriculture residues for fuel applications, Biomass, 21, 115-128. Grenada, E., Lopez-Gonzalez, L. M., Miguez, J. L., Moran, J. (2002). Fuel 573. Husain, Z., Zainac, Z., Abdullah, Z. (2002). Briquetting of palm fiber and shell from pelletizing switchgrass, Website: http://reap-canada.com, accessed on August 26, Lehtikangas, P. (2001). Quality properties of pelletised sawdust, logging residues and bark, Biomass and Bioenergy, 20, 351-360. Mani, S., Tabil, L. G., Sokhansanj, S. (2003a). An overview of compaction of biomass rinds, Powder Handling and Processing, 15, 160-168. ani, S., Tabil, L. G., Sokhansanj, S., Roberge, M. (2003b). Mechanical properties of corn stover grind, ASABE Paper No. 036090, American Society of Agricultural ngineers, Annual International Meeting, July 27-30, Las Vegas, Nevada, USA. ani, S., Tabil, L. G., Sokhansanj, S. (2004a). Grinding performance and physical properties of what and barley straws, corn stover, and switchgrass, Biomass and ioenergy, 27, 339-352. g M E M B 87 Mani, S., Sokhansanj, S., Bi, X., (200 f Corn Stover, Paper No: da. 5). Specific energy requirement for ompacting corn stover, , (Article in Press). sion of straw to high densities in losed cylindrical dies, Journal of Agricultural Engineering Research, 29, 61-72. Lapointe, C. (2000). Assessment of pelletized iofuels, Website: http://reap.ca/Reports/pelltaug2000.html 4b), Compaction o 041160, ASABE/CSAE Annual International Meeting, Ottawa, Ontario, Cana Mani, S., Tabil, L. G., Sokhansanj, S. (200 Bioresource Technologyc O?Dogherty, M. J., Wheeler, J. A. (1984). Compres c Samson, P., Duxbury, P, Drisdelle, M., b , accessed on September, 1, mith, I.E., Probert, S. D., Stokes, R. E., Hansford, R., J. (1977). The briquetting of abil, L.G., Sokhansanj, S. (1996). Process conditions affecting the physical quality of 2004. S wheat straw, Journal of Agricultural Engineering Research, 22, 105-111. T alfalfa pellets, Applied Engineering in Agriculture, 12, 345-350. 88 CHAPTER 4 ? FUTURE WORK The pellets made in the single pellet apparatus have a propensity to expand once fore expansion e calculated using the force-deformation data given as output by the texture analyzer. The density as the probe descends can be calculated by identifying the initial height of e sample in the die before compression begins. At moisture contents below 20.0% there was difficulty in replicating the pellets at given temperature and die size. The difficulty is thought to be due to the placement of e particles in the die. Because of the particle size, and size distribution, the biomass oes not fill the die evenly which can lead to results that can not be easily replicated. maller particles will eliminate filling issues and according to the investigation on article size in Chapter 3, will possibly lead to a more compact pellet. The effect of article size on density and specific energy requirement should be studied on switchgrass ground with 0.79 and 1.59 mm hammer mill screen sizes. COMPACTION BEHAVIOR released from the die. Another way to measure density of the compacts be is of interest. The density of the switchgrass pellets created in the single pellet unit will b th a th d S p p 89 SIMULTANEOUS SACCHARIFICATION AND FERMENTATION (SSF) Large scale ethanol production from biomass can be used for alternative fuel, a fuel additive, or as chemica The use of ethanol as an lternative motor fuel has been increasing throughout the world (Badger, 2002) such as in the United States and Europe (Palm e important s an engine fuel, since it is easily blended with gasoline (Yu and Zhang, 2004). Ethanol can be synthetically made from petroleum or by microbial conversion of biomass materials though fermentation (Badger, 2002). The biomass materials used to produce ethanol fall into three major categories, sugar, starch and cellulosic biomass (Sen, 1989). For example, sugar cane and cane molasses are the basis of the Brazilian fuel ethanol industry (Wayman and Parekh, 1990). Starchy materials commonly used for ethanol production across the globe include, cereal grains (maize and wheat), potato, sweet potato, and cassava, which require hydrolysis to break down the starch into fermentable sugars (Badger, 2002). Currently wheat is being grown as an energy crop and the wheat starch is converted to ethanol in European plants (Palmarola-Adrados, 2005). Because sugar and starch are in the human food chain, and are required for alternative uses they are expensive to use for ethanol production. Therefore, large scale production of ethanol will depend upon the use of a less expensive and renewable feedstock such as lignocellulosics (Seenayya et al., 2000; Badger, 2002). Lignocellulosic materials used for ethanol production originate either as waste materials evolving from processes other than fuel production (agricultural and forest) or as energy crops grown for the purpose of fuel production (Lynd, 1996). Lignocellulosic biomass can provide a resource large enough to be considered as a renewable source of l feed stocks (Seenayya et al., 2000). a arola-Adrados, 2005). Ethanol has becom a 90 liquid transportation fu e ability to product ethanol from lignocellulosic biomass such as corn stover (Kim and Lee 2005b), used recycled paper sludge (RPS), (Lark et al., 1997) used sunflower hulls (Sharma et al., 2004) and switchgrass and poplar (Alizadeh et al., 2005; and Chung, et al., 2005). In general the biomass - to - ethanol conversion process includes three main steps: 1) pretreatment to increase availability/reactivity of the substrate to the enzyme, 2) saccharification (acid or enzymatic) to hydrolyze the cellulose to simple sugars and 3) fermentation to convert the sugars produced during hydrolysis to ethanol (Chung et al., 2005). There are two ways to procure ethanol from lignocellulosics a) separate hydrolysis and fermentation (SHF) or b) simultaneous saccharification and fermentation (SSF). In the SSF process, enzymes and an ethanol producing organism (yeast) is used to carry out simultaneous hydrolysis of cellulose to glucose and the conversion of glucose to ethanol in the same reactor (Lark et al., 1997; Alizadeh et al., 2005). The advantages to the SSF process are: time for ethanol production decreases, less equipment is used, and it has been reported that less enzymes are required (Wyman and Parekh, 1990). The goal for future work is to investigate the difference in ethanol yield between ground and pelleted switchgrass which is of high interest. All the principles and methods stated previously will allow the successful conversion of switchgrass to ethanol. As shown by the previous chapters, pelleting biomass (which has low bulk density) allows simpler handling and decreased transportation to the conversion facility and storage costs before material arrives at the conversion facility. Since the result of the compositional els (Alizadeh et al., 2005). Many researchers have demonstrated th 91 ENZYMATIC HYDROLYSIS Enzymatic hydrolysis is commonly used as a method for breaking down cellulose and hemicellulose into monomeric sugars for fermentation purposes. Compared to acid hydrolysis, enzymatic hydrolysis is milder and more specific. Cellulase is used as the catalyst and is usually carried out at 40-50?C, reducing sugar degradation that occurs at high temperatures and to extend the life of the enzyme (Wyman and Parekh, 1990) and production (Sun and Cheng, 2004). Initial studies using ground and pelleted switchgrass were not obtained. Because this procedure is necessary for successful conversion of lex Dale et al., 1996; Szczodrak and Fiedurek, 1996; Esteghlalian et al., 1997; Soderstrom et analysis between ground and pelleted switchgrass was similar (Chapter 2), it is postulated that the ethanol yield will be as well. Sun and Cheng, 2004). Enzymatic hydrolysis also makes simultaneous saccharification and fermentation feasible making it an extensively studied process for fuel ethanol have already been carried out by the author. Due to equipment malfunction final results switchgrass to ethanol this will be included in future work. PRETREATMENT Pretreatment methods are an important aspect of the bioconversion process. The difficulty in enzymatically hydrolyzing lignocellulosic biomass arises from the comp nature of the material. Cellulose fibrils are embedded in a matrix of lignin and hemicellulose which make the plant tissue resistant to enzymes. Therefore, efficient bioconversion of lignocellulotic feedstock requires pretreatment (Saddler et al., 1993; 92 lignocellulosic biomass increase the specific surface area of the substrate, increase pore volume of the substrate, breakdown the cyrstallinity of the cellulose, and breakdown the lignocellulosic complex, to allow higher ields of fermentable sugars for ethanol production (Saddler, 1993; Dale et al., 1996; Szczodrak and Fiedurek, 1996; Esteghlalian et al., 1997; Kurakake et, al., 2001; Kim and Lee, 2006). Pretreatment effectiveness has been associated with removal of hemicellulose and lignin. Lignin solubilization is beneficial for enzymatic hydrolysis, but the benefits have to be compared with the potential for fermentation inhibition by higher concentration of temperatures (>90?C wet, >160?C dry), researchers have noted that lignin changes upon cooling and does not return to its original form (Lynd, 1996). In developing more effective pretreatments, Cowling and Kirk (1976) suggested placing an emphasis on biological, physical and chemical methods which can alter cyrstallinity and delignify the material which can in turn increase rea of the substrate to the enzyme ore al., 2003; Liu and Wyman, 2005; Alizadeh et al., 2005; Kim and Lee, 2005a; Kim and Lee, 2006). Pretreatment methods for y soluble lignin derivatives. If the lignin does not emerge from the pretreatment stage in the soluble form, it should be chemically modified. Since lignin melts at elevated the available surface a . Other researchers have a contrasting view of altering or decreasing the cyrstallinity where studies have shown examples of effective pretreatments that resulted in unchanged or increased cyrstallinity (Lynd, 1996). The cost of the cellulase enzyme had been a concern in the recent years, theref the most economical and effective pretreatment is desired to decrease enzyme loading (Alizadeh et al., 2005). Industrial economical pretreatment methods include utilizing 93 eed recycle percolation (ARP) process could enable increased digestibility of the swi y was le ore 89) reports that grinding, milling and she acid and alkaline reagents (Kurakake et al., 2001). Typically, hydrolysis yields without pretreatment are < 20% of theoretical yields, where yields after pretreatment can exc 90% of theoretical yields (Lynd, 1996). Initial study on pretreatment methods have already taken place with ground and pelleted switchgrass. Further investigation of pretreatment methods (chemical and mechanical) should take place. For chemical pretreatments, treatment with hot water and the ammonia tchgrass. This is because flow through methods allow lignin and hemicellulose to be removed from the system on a continual basis, which can prevent precipitation of lignin upon cooling and reactions with other components which are present (Kim and Lee, 2006). Also, residence times and temperatures for the soaking with aqueous ammonia (SAA) process should be varied. The switchgrass used in the current stud ground through a hammer mill with a screen size of 3.18 mm. The size of the particles were somewhat large (d gw = 0.865 mm). Reduction of particle size disrupts the crystalline structure and breaks the chemical bonds of the long chain molecules increasing enzyme substrate contact during hydrolysis (Saddler et al., 1987). According to Mani et al (2004a) narrow range particle size distribution with more fines is suitab for enymatic hydrolysis of lignocellulosic materials because of the generation of m surface area and pore spaces during fine grinding. The draw back is that fine grinding of biomass requires high energy consumption. Ladich (19 aring biomass proves to be an effective pretreatment for enzymatic hydrolysis. 94 differ with an alternative calculation method, d Therefore, future research objectives include: 1. Determine how the density values of switchgrass compacts created in a single pellet apparatus will 2. Determine how the density of switchgrass compacts created in a single pellet apparatus will be affected by change in grind size for all temperature, moisture, and die size combinations, 3. Determine the optimal pretreatment methods and conditions (chemical an mechanical) for switchgrass and 4. Investigate the difference in ethanol yield between ground and pelleted switchgrass by using SSF. 95 NCES Alizadeh, H switchgrass by ammonia fiber explosion (AFEX), Applied Biochemistry and Biotechnology Badger, P. and new uses, Edited by Janick, J., and Whipkey A., ASHA Press, Alexandria, VA. Chung, Y., Bakalinsky, A., Penner, M. H. fermentation o ics, Applied Biochemistry and Biotechnology, 121-124, 947-962. Cowling, E. B., Kirk, T. K. Properties of cellulose and lignocellulosic materials as substrates for e Biotechnology and Bioengineering Symposium No. 6, John Wiley and Sons, Inc., 1976, 95-123. Dale, B. E., Leong, C. K., Pham, T. K., Esquivel, V. M., Rios, I., Latimer, V. M. (1996). Hydrolysis of l me levels: application of the AFEX process, Bioresource Technology, 56, 111-116. stghlalian, A., Hashimoto, A. G., Fenske, J. J., Penner, M. H. (1997). Modeling and ptimization of the dilute sulfuric a id pretreatment of corn stover, poplar and itchgrass, Bioresource Technology, 59, 129-136. im, T. H., Lee, Y. Y. (2005a). Pretreatment and fractionation of corn stover by mmonia recycle percolation process, Bioresource Technology, 96, 2007-2013. im, T. H., Lee, Y. Y. (2005b). Pretreatment of corn stover by soaking in aqueous mmonia, Applied Biochemistry and Biotechnology, 121-124, 1119-1132. im, T. H., Lee, Y. Y. (2006). Fractionation of corn stover by hot water and aqueous mmonia treatment, Bioresource Technology, 97, 224-232. urakake M., Kisaka, W., Ouchi, K., Komaki, T. (2001). Pretreatment with ammonia ater for enzymatic hydrolysis of corn husk, bagasse and switchgrass, Applied iochemistry and Biotechnology, 90, 251-259. adisch, M. R. Hydrolysis, In Biomass Handbook, Edited by Kitani, O. and Hall, C. W., ordon and Breach Science ,Amsterdam, 1989, 434-451. ark, N., Xia, Y., Qin, C., Gong, C. S., Tsao, G. T.(1997). Production of ethanol from cycled paper sludge using cellulase and yeast, kluveromyces marxianus, Biomass and ioenergy, 12, 135-143. REFERE ., Teymouri, F., Gilbert, T. I., Dale, B. E. (2005). Pretreatment of , 121-124, 1133-1142. C. (2002). Ethanol from cellulose: a general review, In Trends in new crops (2005). Enzymatic saccharification and f xylose-optimized dilute acid-treated lignocellulos nymatic conversion processes, In ignocellulosics at low enzy E o sw K a K a K a K w B L G L re B 96 Liu, C., Wyman, C.E. (2005). Partial flow of compressed hot water through corn stover enhance hemicellulose sugar recovery and enzymatic digestibility of cellulose, uation of fuel ethanol from cellulosic biomass: chnology, economics, the environment, and policy, Annual Review of Energy and the ani, S., Tabil, L. G., Sokhansanj, S. (2004a). Grinding performance and physical almarola-Adrados, B., Choteborska, P., Galbe M., Zacchi, G. (2005). Ethanol 6, addler, J. N., Chan, M. K. H., Mes-Hartree, M., Breuil, C. Cellulase production and addler, J.N., Ramos, L.P., Breuil, C. Steam Pretreatment of Lignocellulosic residues, In 3). eenayya, G., Reddy, G., Dai Ram, M., Swamy, M. V., Sudha R. K. Production of eview, S., mith, W. H., Science Publishers, Inc., Enfield, New Hampshire, 2000. all, . W., Gordon and Breach Science ,Amsterdam, 1989 ,254-270. tion of enzymatic hydrolysis f sunflower hulls for ethanol production and its scale up, Biomass and Bioenergy, 27, . (2003). Two-step steam pretreatment f softwood by dilute sulfuric acid impregnation for ethanol production, Biomass and un, Y., Cheng, J. (2004). Enzymatic hydrolysis or rye straw and bermuda grass using . (1996). Technology for conversion of lignocellulosic biomass , 10, 367-375. to Bioresource Technology, 96, 1978-1985. Lynd, L. R. (1996). Overview and eval te Environment, 21, 403-465. M properties of what and barley straws, corn stover, and switchgrass, Biomass and Bioenergy, 27, 339-352. P production from non-starch carbohydrates of wheat bran, Bioresource Technology, 9 843-850. S hydrolysis of pretreated lignocellulosic substrates, In Biomass Conversion Technology- Principles and practice, Edited by Moo-Young, M., Pergamon Press (1987), 149-166. S Bioconversion of Forest and Agricultural Plant Residues, Edited by Saddler, J.N. (199 Biotechnology in Agriculture No. 9, 73-92. S ethanol from lignocellulosic materials using Clostridium thermocellum-a critical r In Bioenergy-vision for the new millennium, Edited by Ramamurthi, R., Kastury, S Sen, D. C. Ethanol fermentation, In Biomass Handbook, Edited by Kitani, O. and H C Sharma, S. K., Kalra, K. L., Kocher, G. S. (2004). Fermenta o 399-402. Soderstrom, J., Pilcher, L., Galbe, M., Zacchi, G o Bioenergy, 24, 475-486. S cellulases supplemented with ?-glucosidase, Transactions of the ASABE, 47, 343-349. Szczodrak, J., Fiedurek, J ethanol, Biomass and Bioenergyto 97 ess, Wayman, M., Parekh, S. R. Biotechnology of biomass conversion-fuels and chemicals from renewable resources, Edited by The Institute of Biology, Open University Pr London, 1990. Yu, A., Zhang, H. (2004). Ethanol production of acid-hydrolyzed cellulosic pyrolysate with Saccharomces cerevisiae, Bioresource Technology, 93, 199-204. 98 The quality of pellets obtained from a pellet mill are affected by processing arameters such as moisture content, temperature, particle size, and pellet mill speed. It important to study the effects of these parameters to provide data for the design of iomass handling and processing facilities. The goal of this investigation was to study e compaction behavior of switchgrass. The following specific objectives have been overed throughout the course of this paper : 1. Quantify the effect of moisture content on the physical properties of compacted (pelleted) switchgrass; 2. Evaluate the effect of process parameters on the compaction behavior (density and specific energy) of switchgrass and, 3. Evaluate the effect of pelleting on the composition of switchgrass. It can be concluded from this study that moisture content significantly affected the hysical properties of pellets manufactured from switchgrass. Increasing moisture ontent increased the diameter of the pellets by 8% and decreased the length by 17%. ulk and particle densities decreased by 24 and 16% respectively as moisture content of e pellets increased. A maximum durability rating of 95.91% was obtained when the ellets were at a moisture content of 8.62% (wet basis). Pellets also displayed high to medium durability in the moisture range evaluated. Durability and hardness decreased as a result of increasing moisture. The force required to rupture the pellets ranged from 20.60 to 30.21N. There was a maximum or minimum value of all properties at 8.62% CONCLUSION p is b th c p c B th p 99 moisture content. Relative humidity had a significant effect on the moisture uptake rate constant (k) and equilibrium moisture content (M q ). At lower relative humidity, k was higher. The moisture sorption isotherms showed independence to temperature beyond the relative hum nsity of the switchgrass compacts decreased with There was less relaxation e smaller, and therefore, a denser com samples in different die sizes does not result in e density. In addition the density o required to com he idity of 0.72. For a temperature range between 6 and 50?C the Chung- Pfost equation was the most appropriate model that fit to the experimental data where the Halsey equation was the least appropriate. A relationship between the density and specific energy required and processing parameters (die size, moisture content, and temperature) was established by compressing ground switchgrass using a single pellet apparatus. The de increasing die size resulting from the increase in surface area. of the material as the die size becam pact was formed. Application of the same amount of pressure (95 MPa) to compacts of the sam f the compacts increased as particle size deceased the specific energy press the ground switchgrass decreased with increases in die diameter. For die diameter of 4.8 mm the energy required increased with increasing moisture. T opposite occurred with 6.4 and 7.9 mm die sizes where the specific energy required decreased as moisture content increased. This may be due to the increase in surface area which allows particles to relax more and resistance to compression is lessened. Results from compositional analysis revealed that proportions of individual components for the pelleted samples were higher than those of the ground samples. There were significant differences in ash, acid soluble lignin and acid insoluble lignin and no significant differences in carbohydrates at a 95% significance level. There is no scientific 100 explanation for the increase proportions of individual components of the samples due to pelleting. 101 REFERENCES Alizadeh, H., Teymouri, F., Gilbert, T. I., Dale, B. E. (2005). Pretreatment of switchgrass by ammonia fiber explosion (AFEX), Applied Biochemistry and Biotechnology, 121-124, 1133-1142. Aarseth, K. A., E. Prestlokken, E., (2003). Mechanical properties of feed pellets: Weibull analysis, Biosystems Engineering. 84, 349-361. Adapa, P. K., Schoenau, G. J., Tabil, L. G., Sokhansanj, S., Crerar B.J. (2003). Pelleting of fractionated alfalfa products. ASABE Paper No. 036069: pages 1-11, American Society of Agricultural Engineers Annual international Meeting, July 27-30, Las Vegas, Nevada, USA. Adapa, P. K., Tabil, L. G., Schoenau, G. J., Sokhansanj, S. (2004). Pelleting characteristics of fractionated sun-cured and dehydrated alfalfa grinds, Applied Engineering in Agriculture, 20, 813-820. ASABE (2001). Moisture relationship of plant based agricultural products, ASABE Standard D245.5, St. Joseph, MI. ASABE (2002). Cubes, pellets, and crumbles--definitions and methods for determining density, durability, ASABE Standard, S269.4, St. Joseph, MI. ASABE (2003). Method of determining and expressing fineness of feed materials by sieving, ASABE Standard, S319.3, St. Joseph, MI. Badger, P. C. (2002). Ethanol from cellulose: a general review, In Trends in new crops and new uses, Edited by Janick, J., and Whipkey A., ASHA Press, Alexandria, VA. Badger P.C. Biomass Transport Systems. In: Encyclopedia of Agriculture, Food and Biological Engineering, Edited by Heldman, D., New York: Marcel Dekker Publisher; 2003; 94-98 Balasubramanian, D. (2001). Physical properties of raw cashew nut, Journal of Agricultural Engineering Research, 78, 291-297. Barker, G. L. (1996). Equilibrium moisture content of cotton plant components, Journal of Agriculture Engineering Research, 63, 353-364. 102 Bellinger, P. L., McColly, H. F. (1961). Energy requirements for forming hay pellets, Agriculture Engineering, 42, 244-247. Bello, M. O., Tolaba, M. P., Suarez, C. (2005). Water absorption and starch gelatinization in whole rice grain during soaking, LWT ? Food Science and Technology, In Press, Corrected Proof, Available online November 2005. 000). Switchgrass co-firing: pilot scale and filed ting the pelleting of hay, Agriculture acteristics of some biomass briquettes prepared s, N. E. (2002). The yield and composition of . mperature and relative humidity during storage, An ASABE/CSAE Meeting tawa, : application of the AFEX process, Bioresource Technology, 56, 111-116. Bruhn, H.D., Mechanical treatment, In Biomass and Handbook, ed. Kitani,O. and Hall., C.W., Gordon and Breach Science, Amsterdam, 1989, 452-459. Boylan, D., Bush, V., Bransby, D. I. (2 evaluation, Biomass and Bioenergy, 19, 411-417. Butler, J. L., McColly H. F. (1959). Factors affec Engineering, 40, 442-446. Chandrasekar, V., Viswanathan, R. (1999). Physical and thermal properties of coffee, Journal of Agriculture Engineering Research 73, 227-234. Chhinnan, M. S., (1984), Evaluation of selected mathematical models for describing thin layer drying of in-shell pecans, Transactions of the ASABE, 27,610-615. Chin, O. C., Siddiqui, K. M. (2000). Char under modest die pressures, Biomass and Bioenergy, 18, 223-228. Christian, D. G., Riche, A. B., Yate switchgrass and coastal panic grass grown as a biofuel in Southern England, Bioresource Technology, 83, 115-124. Chung, Y., Bakalinsky, A., Penner, M. H. (2005). Enzymatic saccharification and fermentation of xylose-optimized dilute acid-treated lignocellulosics, Applied Biochemistry and Biotechnology, 121-124, 947-962 Cowling, E. B., Kirk, T. K. Properties of cellulose and lignocellulosic materials as substrates for enymatic conversion processes, In Biotechnology and Bioengineering Symposium No. 6, John Wiley and Sons, Inc., 1976, 95-123. Dadgar, S., Tabil, L. G., Crerar, B. (2004). Moisture Adsorption of files peas as a function of te Presentation, Paper NO. 046089, American Society of Agricultural Engineers/Canadien Society of Agricultural engineers, Annual International Meeting, August 1-4, Ot Ontario, Canada. Dale, B. E., Leong, C. K., Pham, T. K., Esquivel, V. M., Rios, I., Latimer, V. M. (1996). Hydrolysis of lignocellulosics at low enzyme levels 103 f 89-98. ontent of characteristics of rbas, M., Ertugay, M. F., Certel, M. (2005). Moisture absorption behavior of semolina stghlalian, A., Hashimoto, A. G., Fenske, J. J., Penner, M. H. (1997). Modeling and chnology, 59, 129-136. 51-63. alfalfa g m the : Deshpande, S. D., Bal, S., Ojha, T. P. (1993). Physical properties of soybean, Journal o Agricultural Engineering Research, 56, Dexter, S. T., Sheldon, W. H., Waldron, D. I. (1947). Equilibrium Moisture c alfalfa hay, Agricultural Engineering, 28, 295-296. Durkova, A. G., N.D. Menkov N. D. (2005). Moisture sorption chickpea flour, Journal of Food Engineering, 68, 535-539. E and farina, Journal of Food Engineering, 96, 191-198. E optimization of the dilute sulfuric a id pretreatment of corn stover, poplar and switchgrass, Bioresource Te Faborode M. O. (1990). Analysis of extrusion compaction of fibrous agricultural residues for fuel applications, Biomass, 21, 115-128. Fasina, O.O., Sokhansanj, S. (1992). Hygroscopic moisture absorption by alfalfa cubes and pellets, Transactions of the ASABE, 35, 1615-1619. Fasina, O.O., Sokhansanj, S. (1993). Equilibrium moisture relations and hat of sorption of alfalfa pellets, Journal of Agricultural Engineering Research, 56, Fasina, O.O., Sokhansanj, S., Irudayaraj, J. (1994). Thermo-physical properties of pellets, Paper No. 946022, Annual International Meeting of ASABE, Kansas City, MO, June 19-22. Fasina O.O., Sokhansanj, S. (1996). Storage and handling characteristics of alfalfa pellets, Powder Handling and Processing, 8, 361-365. Gilpin, A. S., Herrman, T. J., Behnke, K. C., Fairchild, F. J. (2002). Feed moisture, retention time, and steam as quality and energy utilization determinants in the pelletin process, Applied Engineering in Agriculture, 18, 331-338. Husain, Z., Zainac, Z., Abdullah, Z. (2002). Briquetting of palm fiber and shell fro processing of palm nuts to palm oil, Biomass and Bioenergy, 22, 505-509. Jannasch, R., Samson, R., de Maio, A., Adams, T. (2001a). Changing the energy climate clean and green heat from grass biofuel pellets, Website: http://reap?canada.com, ccessed on Augua st 27, 2004. Jannasch, R., Quan, Y., Samson, R. (2001b). A process and energy analysis of pelletizing switchgrass, Website: http://reap-canada.com, accessed on August 26, 2004. 104 iofuel pellets, Website: http://reap- Jannasch, R., Samson, R., de Maio, A., Adams, T., Ho Lem, C. (2001c). Changing the energy climate: clean and green heat from grass b canada.com, accessed on August 27, 2004. mass Handbook, Editors: Kitani, a, S. N. (1999). Physical and hygroscopic properties of makhana, Journal of shi, D. C., Das, S. K., Mukherjee, R. K. (1993). Physical properties of pumpkins seeds, ctionation of corn stover by mmonia recycle percolation process, Bioresource Technology, 96, 2007-2013. stover by soaking in aqueous mmonia, Applied Biochemistry and Biotechnology, 121-124, 1119-1132. er by hot water and aqueous mmonia treatment, Bioresource Technology, 97, 224-232. a cubes uring shipping and storage, Applied Engineering in Agriculture, 15, 671-676. ymatic hydrolysis of corn husk, bagasse and switchgrass, Applied iochemistry and Biotechnology, 90, 251-259. easurement and use, The merican Association of Cereal Chemists, ?1984. n stover Jenkins, B.M., Physical Properties of Biomass, In Bio O., Hall, C.W., Gordon and Breach Science, Amsterdam, 1989, 860 ? 891. Jh Agricultural Engineering Research, 72, 145-150. Jo Journal of Agricultural Engineering Research, 54, 219-229. Kim, T. H., Lee, Y. Y. (2005a). Pretreatment and fra a Kim, T. H., Lee, Y. Y. (2005b). Pretreatment of corn a Kim, T. H., Lee, Y. Y. (2006). Fractionation of corn stov a Khoshtagaza, M. H., Sokhansanj, S., Gossen, B. D. (1999). Quality of alfalf d Kurakake M., Kisaka, W., Ouchi, K., Komaki, T. (2001). Pretreatment with ammonia water for enz B Labuza, T. P., Moisture sorption: aspects of isotherm m A Ladisch, M. R., Hydrolysis, In Biomass Handbook, Edited by Kitani, O. and Hall, C. W., Gordon and Breach Science ,Amsterdam, 1989, 434-451. Lark, N., Xia, Y., Qin, C., Gong, C. S., Tsao, G. T. (1997). Production of ethanol from recycled paper sludge using cellulase and yeast, kluveromyces marxianus, Biomass and ioenergy, 12, 135-143. B Lehtikangas, P. (2001). Quality properties of pelletised sawdust, logging residues and bark, Biomass and Bioenergy, 20, 351-360. Liu, C., Wyman, C.E. (2005). Partial flow of compressed hot water through cor to enhance hemicellulose sugar recovery and enzymatic digestibility of cellulose, ioresource Technology, 96, 1978-1985. B 105 e cizing effect, arbohydrate Polymers, 60, 283-289. overview of compaction of biomass rinds, Powder Handling and Processing, 15, 160-168. Mechanical properties of orn stover grind, ASABE Paper No. 036090, American Society of Agricultural ani, S., Tabil, L. G., Sokhansanj, S. (2004a). Grinding performance and physical a, Ontario, Canada. D., ng switchgrass as a bioenergy crop, : Perspectives on new crops and new uses, Edited by Janick J., Alexandria, VA: ASHS cLaughlin, S.B., Kszos, L.A. (2005). Development of switchgrass (Panicum virgatum) , S.A., Montross, M.D. (2004). Effect of moisture content and Lynd, L. R. (1996). Overview and evaluation of fuel ethanol from cellulosic biomass: technology, economics, the environment, and policy, Annual Review of Energy and th Environment, 21, 403-465. Mali, S., Sakanaka, L. S., Yamashita, F., Grossmann, M. V. E. (2005). Water sorption and mechanical properties of cassava starch films and their relation to plasti C Mani, S., Tabil, L. G., Sokhansanj, S. (2003a). An g Mani, S., Tabil., L. G., Sokhansanj, S., Roberge, M., (2003b), c Engineers, Annual International Meeting, July 27-30, Las Vegas, Nevada, USA. M properties of what and barley straws, corn stover, and switchgrass, Biomass and Bioenergy, 27, 339-352. Mani, S., Sokhansanj, S., Bi, X. (2004b). Compaction of Corn Stover, Paper No: 041160, ASABE/CSAE Annual International Meeting, Ottaw Mani, S., Tabil, L. G., Sokhansanj, S. (2005). Specific energy requirement for compacting corn stover, Bioresource Technology, (Article in Press). McMullen, J., Fasina, O.O., Wood, C. W., Feng, Y. (2005). Storage and handling characteristics of pellets from poultry litter, Applied Engineering in Agriculture, 21, 645- 651. McLaughlin, S.B., Bouton, J., Bransby, D., Conger, B., Ocumpaugh, W., Parrish, Taliaferro, C., Vogel, K. Wullschleger, S. Developi In Press; 1999, p. 282?299. M as a bioenergy feedstock in the United States, Biomass and Bioenergy, 24, 515-535. cNeill, S.G., ThompsonM broken kernels on the bulk density and packing of corn, Applied Engineering in Agriculture, 20, 475-480. Missaoui, A., Fasoula, V., Bouton, J. (2005). The effect of low plant density on response to selection for biomass production in switchgrass, Euphytica, 142: 1-12. 106 , imkar, P.M., Chattopadhyay, P. K. (2001). Some physical properties of green gram, REL (2004). Laboratory Analytical Procedures - Determination of Structural REL (2005). Laboratory Analytical Procedure - Determination of ash in biomass, ler, J. A. (1984). Compression of straw to high densities in losed cylindrical dies, Journal of Agricultural Engineering Research, 29, 61-72. , ournal of Food Engineering, 67, 441-450. . Ethanol roduction from non-starch carbohydrates of wheat bran, Bioresource Technology, 96, ka, M., (2004), Global biomass fuel resources, Biomass and Bioenergy, 27, 613- 20. annual supply, ebsite: http://feedstockreview.ornl.gov/pdf/billion_ton_vision.pdf Mohsenin, N. N. (1986). Physical Properties of Plant and Animal Materials: Structure Physical Characteristics and Mechanical Properties (2 nd Ed). New York: Gordon and Breach Publishers. Nelson, S. O. (2002). Dimensional and density data for seeds of cereal grains and other crops, Transactions of the ASABE, 45, 165-170. N Journal of Agricultural Engineering Research, 80, 183-189. N Carbohydrates and Lignin in Biomass, United States Department of Energy, Washington D.C. N United States Department of Energy, Washington D.C. O?Dogherty, M. J., Whee c Pagano, A. M., Mascheroni, R. H. (2005). Sorption isotherms for amaranth grains J Palmarola-Adrados, B., Choteborska, P., Galbe M., Zacchi, G. (2005) p 843-850. Parik 6 Perlack, R. D., Wright, L. L., Turhollow, A. F., Grahm, R. L. Stokes, B. J., Erbach, D. C. (2005). Biomass as feedstock for a bioenergy and bioproducts industry: the technical feasibility of a billion ton W , accessed on oth, A. M., Sample, D. W., Ribic, C. A., Paine, L., Undersander, D. J., Bartelt, G. A. . H., Mes-Hartree, M., Breuil, C., Cellulase production and December 15, 2005. R (2005) Grassland bird response to harvesting switchgrass as a biomass energy crop, Biomass and Bioenergy, 28, 490-498. addler, J. N., Chan, M. KS hydrolysis of pretreated lignocellulosic substrates, In Biomass Conversion Technology- Principles and practice, Edited by Moo-Young, M., Pergamon Press (1987), 149-166. 107 gy in Agriculture No. 9, 73-92. , A. A., Ocumpaugh, W. R., Hussey, . A., Read, J. C., Tischler, C. R. (1996). Switchgrass as a sustainable bioenergy crop, anderson, M. A., Egg, R. P., Wiselogel, A. E. (1997). Biomass losses during harvest nte, C. (2000). Assessment of pelletised iofuels, Website: http://reap.ca/Reports/pelltaug2000.html Saddler, J.N., Ramos, L.P., Breuil, C., Steam Pretreatment of Lignocellulosic residues, In Bioconversion of Forest and Agricultural Plant Residues, Edited by Saddler, J.N., (1993), Biotechnolo Sanderson, M. A., Reed, R. L., McLaughlin, S. D., Wullschleger, S. D., Conger, B. V., Parrish D. J., Wolf, D. D., Taliaferro, C., Hopkins M Bioresource Technology, 56, 83-93. S and storage of switchgrass, Biomass and Bioenergy, 12, 107-114. Samson, P., Duxbury, P., Drisdelle, M., Lapoi b , accessed on August 16, AS (2002-2003). SAS Users? Guide: Statistics: SAS Institute Inc., Cary NC. f thanol from lignocellulosic materials using Clostridium thermocellum-a critical review, ld, New Hampshire, 2000. Hall, C. harma, N., Piscioneri, I., Pignatelli, V. (2003). An evaluation of biomass yield stability 53-2958 quilibruim moisture content of biomass briquettes, Biomass and ionenergy, 26, 251-253. gineering Research, 22, 105-111. t 2004. S Seenayya, G., Reddy, G., Dai Ram, M., Swamy, M. V., Sudha R. K., Production o e In Bioenergy-vision for the new millennium, Edited by Ramamurthi, R., Kastury, S., Smith, W. H., Science Publishers, Inc., Enfie Sen, D. C, Ethanol fermentation, In Biomass Handbook, Edited by Kitani, O. and W., Gordon and Breach Science ,Amsterdam, 1989 ,254-270. S of switchgrass (Panicum virgatum L.) cultivars, Energy Conversion and Management, 44, 29 Sharma, S. K., Kalra, K. L., Kocher, G. S. (2004). Fermentation of enzymatic hydrolysis of sunflower hulls for ethanol production and its scale up, Biomass and Bioenergy, 27, 399-402. Singh, R.N. (2004). E B Smith, I.E., Probert, S. D., Stokes, R. E., Hansford, R., J. (1977). The briquetting of wheat straw, Journal of Agricultural En Soderstrom, J., Pilcher, L., Galbe, M., Zacchi, G. (2003). Two-step steam pretreatmen of softwood by dilute sulfuric acid impregnation for ethanol production, Biomass and Bioenergy, 24, 475-486. 108 oysal, Y., Oztekin, S. (2001). Comparison of seven equilibrium moisture content hydrolysis or rye straw and bermuda grass using ellulases supplemented with ?-glucosidase, Transactions of the ASABE, 47, 343-349. gnocellulosic biomass ethanol, Biomass and Bioenergy, 10, 367-375. of ellets, Applied Engineering in Agriculture, 12, 345-350. o its ompaction characteristics, Applied Engineering in Agriculture, 13, 499-505. homas, M., A. F. B. van der Poel (1996). Physical quality of pelleted animal feed 1. an den Berg, S. Bruin, Water activity and its estimation in food system: theoretical 004). Pellets 2002: the first world conference on pellets, Biomass and ioenergy, 27, 513-520. ettes as possible fuels for Kenya, Biomass and Bioenergy, 8, 175-179. ress, ondon, 1990. ., Sanderson, . A. (1996). Compositional changes during storage of large round switchgrass bales, son, J. S., Bulk solids handling: an introduction to the practice nd technology. Leonard Hill, New York, 1987. Sokhansanj, S., Turhollow A. F. (2004). Biomass densification ?cubing operations and costs for corn stover, Applied Engineering in Agriculture, 20, 495-499. S equations for some medicinal and aromatic plants, Journal of Agriculture Engineering Research 78, 57-63. Sun, Y., Cheng, J. (2004). Enzymatic c Szczodrak, J., Fiedurek, J. (1996). Technology for conversion of li to Tabil, L. G., Sokhansanj, S. (1996). Process conditions affecting the physical quality alfalfa p Tabil, L. G., Sokhansanj S. (1997). Bulk properties of alfalfa grind in relation t c Temple, S. J., van Boxtel, A. J. B. (1999). Equilibrium moisture content of tea, Journal of Agricultural Engineering Research, 74, 83-89. T Criteria for pellet quality, Animal Feed Science Technology, 61, 89-112. v aspects, Editors: L. B. Rockand, G. F. Stewart. Academic Press Inc. New York. 1981. Vinterback, J. (2 B Wamukonya, L., Jenkins, B. (1995). Durability and relaxation of sawdust and wheat- straw briqu Wayman, M., Parekh, S. R., Biotechnology of biomass conversion-fuels and chemicals from renewable resources, Edited by The Institute of Biology, Open University P L Wiselogel, A. E., Agblevor, F. A., Johnson, D. K., Deutch, S., Fennell, J. A M Bioresource Technology, 56, 103-109. Woodcock, C. R. and Ma a 109 01-205. ate ink, F. J. (1935). Equilibrium moistures of some hays, Agricultural Engineering, 16,, han, B. F., Chen, X., Noon, C. E., Wu, G. (2005). A GIS-enabled comparison of fixed 295-306. Yang, W., Sokhansanj, S., Crerar, W. J., Rohani, S. (1996). Size and shape related characteristics of alfalfa grind, Canadian Agricultural Engineering, 38, 2 Yu, A., Zhang, H. (2004). Ethanol production of acid-hydrolyzed cellulosic pyrolys with Saccharomces cerevisiae, Bioresource Technology, 93, 199-204. Z 451-452. Z and discriminatory pricing strategies for potential switchgrass ? to ? ethanol conversion facilities in Alabama, Biomass and Bioenergy, 28, 110 APPENDICES 111 APPENDIX A Table A.1 ? Physical properties of switchgrass pellets (sample sizes are in parentheses). Moisture Content (%, w.b.) m) (50) Stdev (mm) Diameter (m 6.32 4.85 0.05 8.62 4.84 0.06 11.00 4.90 0.07 14.84 5.02 0.11 17.00 5.25 0.20 Moisture Content (%, w.b.) Length (mm) (50) Stdev (mm) 6.32 33.61 3.28 8.62 35.27 3.91 11.00 32.80 3.46 14.84 32.35 3.80 17.00 27.97 4.03 Moisture Content (%, w.b.) Particle Density (kg/m 3 ) (3) Stdev (kg/m 3 ) 6.32 1451.27 6.97 8.62 1462.37 1.07 11.00 1455.99 5.07 14.84 1445.13 5.80 17.00 1434.38 2.86 Moisture Content (%, w.b.) Bulk Density (kg/m 3 ) (2) Stdev (kg/m 3 ) 6.32 683.61 0.39 8.62 707.62 0.76 11.00 673.79 6.03 14.84 612.45 2.29 17.00 536.21 4.57 Moisture Content (%, w.b.) Durability (%) (3) Stdev (%) 6.32 95.91 0.31 8.62 96.65 0.43 11.00 94.43 0.19 14.84 89.06 0.53 17.00 78.44 0.67 Moisture Content (%, w.b.) Hardness (N) (50) Stdev (N) 6.32 30.04 12.21 8.62 30.21 9.36 11.00 26.03 9.58 14.84 24.79 10.13 17.00 21.60 7.84 112 6 7 8 9 12 0 5 10 15 25 30 35 Time tu r e C o n t e n 10 t (% , 11 b . 20 (h) M o i s w . ) 15?C 25?C 35?C 45?C Predicted Figure A.1- Moisture change in switchgrass pellets exposed to air at 65% relative ty with varying temperatures (15, 25, 35, and 45?C). moistur humidi Initial e content was 7.19% (w.b.). 6 8 9 11 12 0 1020 4050 Time t u re C o n t en t , w . b . 7 Moi s 10 ( % 30 (h) ) 15?C 25?C 35?C Predicted Figure A.2 ? Moisture change in switchgrass exposed to air at 80 tive idity with v (15, 25 and 35 pellets % rela hum arying temperatures ?C). 6 7 8 9 10 11 12 0 5 10 15 20 25 30 35 Time (h) M o i s tu r e C o n te n t (% , w . b . ) 50% 65% 80% Predicted tempe 0%). Figure A.3 - Moisture change in switchgrass pellets exposed to air at constant rature of 65 and 8 25?C with varying relative humidities (50, 6 7 8 9 10 11 12 0 5 10 15 20 25 30 35 ) M o i s tu r e C o n te n t (% , w . b . 50% 65% 80% Predicted Time (h) Figure A.4 - Moistur d to air at constant temperature of 35?C with idities (50, 65 and 80%). e change in switchgrass pellets expose varying relative hum 113 APPENDIX B Table B.1 ? Density of switchgrass pellets at 10.4 and 13.2% moisture content (w.b.). Die Size Temp (?C) Moisture (%, w.b.) Density (kg/m 3 ) 4.8 mm 60 10.4 1122.98 75 10.4 1206.76 90 10.4 1215.88 6.4 mm 60 10.4 1036.61 75 10.4 1052.09 90 10.4 1023.30 7.9 mm 60 10.4 966.09 75 10.4 976.16 90 10.4 923.31 4.8 mm 60 13.2 1122.07 75 13.2 1151.63 90 13.2 1206.31 6.4 mm 60 13.2 1023.30 75 13.2 1098.42 90 13.2 1024.27 7.9 mm 60 13.2 956.76 75 13.2 960.56 90 13.2 1006.86 114 115 Table B.2 ? Density of switchgrass pellets at 16.2 and 20.0% moisture content (w.b.). Die Size Temp (?C) Moisture (%, w.b.) Density (kg/m 3 ) 4.8 mm 60 16.2 1049.05 75 16.2 1048.33 90 16.2 1075.72 6.4 mm 60 16.2 998.54 75 16.2 936.08 90 7.9 mm 60 16.2 1045.93 16.2 861.37 7 9 4.8 mm 6 5 16.2 898.80 0 16.2 946.78 0 20.0 750.68 7 9 6.4 mm 6 5 20.0 862.28 0 20.0 955.21 0 20.0 784.98 7 9 7.9 mm 6 5 20.0 728.61 0 20.0 904.66 0 20.0 714.73 7 9 5 20.0 732.06 0 20.0 844.37 116 a) 6 800 12 60 75 90 Temp (?C) D e n s i t y ( k g/ m 3 ) 00 1000 00 1400 erture 4.8mm 6.4mm 7.9mm b) 600 800 1200 60 Tempe (?C) D e ns i t y ( k g/ m 3 ) 1000 1400 75 90 rature 4.8mm 6.4mm 7.9mm Figure B.1 ? Density of switchgrass pellets ed by temperature and die size at constant mo re contents of a) and b) 16.2% (w affect 13.2%istu .b.). a) 600 800 1000 1200 1400 4.8mm 6.4mm 7.9mm Die Size D e n s i t y ( k g/ m 3 ) 10.4% 13.2% 16.2% 20.0% b) 600 800 1000 1200 4.8mm 6.4mm 7.9mm Die Size D e ns i t y ( k g/ m 3 ) 1400 10.4% 13.2% 16.2% 20.0% c) 600 800 1000 1400 4.8mm 6.4mm 7.9mm Die Size D e n s i t y ( k g/ m 1200 3 ) 10.4% 13.2% 16.2% 20.0% Figure B.2 ? Density of pellets at constant temperature of a) 60?C b) 75?C c) 90?C. 117 118 a) 600 800 1000 1200 1400 60 75 90 4.8mm 4.8mm 4.8mm Temperature (?C) D e ns i t y ( k g/ m 3 ) 10.4% 13.2% 20.0% b ) 600 800 1000 1200 1400 60 75 90 6.4mm 6.4mm 6.4mm Temperature (?C) D e n s i t y ( k g/ m 3 ) 10.4% 13.2% 20.0% c ) 600 800 1000 1200 1400 60 75 90 7.9mm 7.9mm 7.9mm Temperature (?C) D e ns i t y ( k g/ m 3 ) 10.4% 13.2% 20.0% F . igure B.3 ? Density of pellets at constant die sizes of a) 4.8 mm b) 6.4 mm c) 7.9 mm Table B.3 ? Specific energy used to compress ground switchgrass at 10.4 and 13.2% moisture content (w.b.). Die Size Temp (?C) Moisture (%, w.b.) Specific Energy (MJ/t) 4.8mm 60 10.4 75.18 75 10.4 54.42 90 10.4 47.75 6.4mm 60 10.4 70.51 75 10.4 57.92 90 10.4 82.07 7.9mm 60 10.4 29.26 75 10.4 23.71 90 10.4 26.92 4.8mm 60 13.2 62.04 75 13.2 61.32 90 13.2 44.34 6.4mm 60 13.2 22.48 75 13.2 47.17 90 13.2 65.27 7.9mm 60 13.2 25.24 75 13.2 21.25 90 13.2 24.15 119 120 Die T ( y t) Table B.4 ? Specific energy used to compress ground switchgrass at 16.2 and 20.0% moisture content (w.b.) Size emp ?C) Moisture (%, w.b.) Specific Energ (MJ/ 4.8 6 1 mm 0 16.2 64.4 7 8 9 5 6.4 6 2 5 16.2 59.5 0 16.2 81.3 mm 0 16.2 60.1 7 7 9 8 7.9 6 0 5 16.2 76.9 0 16.2 72.1 mm 0 16.2 23.0 7 6 9 1 4.8 6 1 5 16.2 57.4 0 16.2 20.8 mm 0 20.0 88.8 7 3 9 8 6.4 6 6 5 20.0 63.8 0 20.0 54.9 mm 0 20.0 53.4 7 4 7 7 7.9 6 2 5 20.0 48.9 5 20.0 42.5 mm 0 20.0 23.7 75 90 20.0 22.48 20.0 19.33 Note: Unless specified all experiments were preformed in duplicate and mean values are reported.