fir 5%. 3.11%. i . V N . hwmrxh—rmuk J3. mm in _. - “IT... _ .3. ... fishy. ... .L. x _ p . .5: fix 4 . Span , .. .5 ‘ v . 2.3 . :4: UNIV?! . viva . 1 bw ‘ 33%. a.» z gm .V, Fan «hum. w»... .3. 1 . 19ft. hut, . . Edwam . 1.: ”VF, I: {:Icll gram .[.h. : 1—1 _..fl|...wn(, ~1N~tVI>mZ1 ‘ issAn . . .11 3.20.. l 2,... ,a‘.., LIIRARY I' Michigan State University 1 This is to certify that the thesis entitled PREDICTING DIGESTIBILITY OF AMMONIA FIBER EXPLOSION (AFEX) TREATED RICE STRAW presented by LISA ELI SHEBA GOLLAPALLI has been accepted towards fulfillment of the requirements for MASTERS degree in CHEMICAL ENGINEERING Date 5! H (0‘ W1 M Major professor 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution PLACE IN RETURN BOX to remove this checkout from your record. To AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE .1 011310 2E 2005 t ”le 2007 000013 a 2.002 6/01 clelRC/DateDuepes-sz PREDICTING DIGESTIBILITY OF AMIVIONIA FIBER EXPLOSION (APEX) TREATED RICE STRAW By Lisa Elisheba Gollapalli A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Chemical Engineering 2001 ABSTRACT PREDICTIN G DIGESTIBILITY OF AMMONIA FIBER EXPLOSION (APEX) TREATED RICE STRAW By Lisa Elisheba Gollapalli We modeledthe enzymatic digestibility of the Ammonia Fiber Explosion (AFEX) treated rice straw by statistically correlating the variability of the samples to differences in treatment using several different analytical techniques. Li gnin content and crystallinity of cellulose affect enzymatic hydrolysis the most. X-ray diffraction was used to measure the crystallinity index while fluorescence and diffuse reflectance infrared (DRIFT) spectroscopy measured the lignin content of the samples. Multivariate analysis was applied to correlate the sugar yields of the various samples with x-ray diffraction and spectroscopic data. Principal Component Analysis (PCA) and multi-linear regression (MLR) techniques could not predict the digestibility of the rice straw samples. The best correlation (R2 value of 0.6) was found between the initial treatment conditions of the APEX process and the concentration of xylose at 24 hours after enzymatic hydrolysis. DEDICATION I would like to thank my Lord and Savior Jesus Christ for the wisdom, strength and ability He granted to me to accomplish my thesis work. I would also like to thank my family, friends and The Little Flock Christian Fellowship for their prayer and support during my Masters program in the Department of Chemical Engineering at Michigan State University. The following is a verse from the Bible, which sustained me through all the hard work. “ My grace is sufficient for thee, for My strength is made perfect in weakness." - 2 Corinthians 12:19 iii ACKNOWLEGEMENTS The author wishes to thank Dr. Bruce Dale for his support and guidance throughout the course of this project; Dr. Elankovan Poonampalam and Dr. Doug Rivers for their suggestions and comments; Denise Rummler and Mike Guettler for technical assistance in the hydrolysis of rice straw; Dr. Gilliland for helping me understand the statistics analysis and Arun Ross for helping me with Matlab and other statistical packages. iv 1 INTRODUCTION 1.1 Treatments 1.2 Ammonia Fiber Explosion (AFEX) 1.3 Problem 2 LITERATURE REVIEW 2.1 Hypothesis 2.2 Technique to Determine Crystallinity Index (CrI) 2.3 Technique to determine Lignin Content 2.4 Problem Solving Approach 3 HYDROLYSIS 3.1 Experimental 3.1.1 3.1.2 3.1.3 3.1.4 3.1.5 3.1.6 3.1.7 3.2 3.2.1 3.2.2 3.2.3 TABLE OF CONTENTS Delignification of Rice Straw Determining moisture content of rice straw High Pressure Liquid Chromatography (HPLC) DIONEX Acid Hydrolysis Enzymatic Hydrolysis Statistical Modeling Results and Discussion Factors affecting Enzymatic Hydrolysis Composition of rice straw Statistical Correlation ESBK‘GEEEEESS b) N .5. UI 3.3 Summary 55 4 CRYSYTALLINITY INDEX 57 4.1 Introduction 57 4.2 Experimental 58 4.2.1 Methods 58 4.3 Results and Discussion 59 4.3.1 Trends in Cd 59 4.32 Statistical analysis 69 4.4 Summary 77 5 FLUORESCENCE 79 5.1 Introduction 79 5.2 Experimental 80 5.2.1 Methods 5.2.2 Statistical analysis methods 82 5.3 Results and Discussion s 5.3.1 Reproducibility of Fluorescence Profiles 5.3.2 Trends in Fluorescence Spectra 21° 91 5.3.3 Statistical Analysis 5.4 Summary 6 DRIFT 6.1 Introduction 6.2 Experimental 6.2.1 DRIFT Spectroscopy Method vi 100 102 102 103 103 6.2.2 Principal Component Regression 104 6.3 Results and Discussion 104 6.3.1 Spectroscopic Characterization of AFEX Treated Rice Straw............. 104 6.3.2 Multivariate Analysis . 112 6.4 Summary 114 7 CONCLUSION 115 7.1 Recommendations ' 117 8 APPENDIX 118 9 . BIBLIOGRAPHY 134 vii LIST OF TABLES Table 3.1 AFEX Process Parameters for Treated Rice Straw 21 Table 3.2 Moisture Content of the AFEX Treated Rice Straw 30 Table 3.3 R2 values for Correlation between Treatment Conditions and Sugar ...... 47 Table 4.1 Crystallinity Index of the AFEX Treated Rice Straw 60 Table 4.2 R2 Values for Correlation between Cd and Sugar 76 Table 5.1 Fluorescence data 94 Table 5.2 R2 values from MLR Analysis of Fluorescence data and Cone. of sugar.. 96 Table 5.3 R2 values from MLR Analysis of Fluorescence data and Percent Yield of Sugar 97 Table 5.4 R2 values from PCR analysis of fluorescence data and sugar ................. 98 Table 5.5 R2 Values from MLR Analysis of Fluorescence Data and Treatment Conditions 100 Table 6.1 R2 values for PCR analysis between DRIFT spectra and sugar.............. 113 Table 8.1Concentration of Sugar at 3hr after Enzymatic Digestion 119 Table 8.2 Concentration of Sugar at 24hr Enzymatic Digestion 121 Table 8.3 Concentration of Sugar at 48hr Enzymatic Digestion 123 Table 8.4 Percent Yield of Sugar at 3hr Enzymatic Digestion 125 Table 8.5 Percent Yield of Sugar at 24hr after Enzymatic Digestion 127 Table 8.6 Percent Yield of Sugar at 48hr Enzymatic Digestion 129 Table 8.7 MatLab Code for Principal Component Analysis 131 viii LIST OF FIGURES Figure 2.1 Schematic View of Plant Cell Structure 8 Figure 2.2 Precursors of Guaiacyl, p-hydroxypbenol and syringyl Lignin ............... 9 Figure 2.3 Repeating Unit of Cellulose Polymer Figure 3.1 Sugars from Hydrolysis Studies Figure 3.2 Trends in the Concentration of Dextrose Figure 3.3 Trends in the Concentration of Xylose Figure 3.4 Trends in Percent Yield of Dextrose Figure 3.5 Trends in Percent Yield of Xylose Figure 3.6 Predicting Ultimate Yield of Dextrose using Initial Hydrolysis Rate ...... Figure 3.7 Predicting Ultimate Yield of Xylose using Initial Hydrolysis Rate ......... Figure 3.8 Predicting Ultimate Yield of Total Sugar using Initial Hydrolysis Rate . Figure 4.1 X-ray Diffraction profile of Sample B108 Figure 4.2 X-Ray Diffraction Profile of sample C172 Figure 4.3 X-ray Diffraction profile for sample B66 Figure 4.4 X-ray Diffraction profile of untreated rice straw Figure 4.5 Cd versus concentration of dextrose at 3hr. Figure 4.6 Cd versus concentration of Dextrose at 24hr Figure 4.7 Cd versus concentration of dextrose at 48hr Figure 4.8 Cd versus concentration of xylose at 3hr Figure 4.9 Cd versus concentration of Xylose at 24hr - Figure 4.10 CrI versus concentration of xylose at 48hr Figure 5.1 Effect of Delignification on Fluorescence Spectra ix 10 33 38 39 43 44 51 52 65 67 70 71 72 73 74 75 86 Figure 5.2 Fluorescence Profile for Powdered (106um) Rice Straw 88 Figure 5.3 Fluorescence Spectra Reproducibility of 0.35mm Size Rice Straw ........ 89 Figure 5.4 Effect of Particle Size on the Fluorescence Spectra 90 Figure 5.5 Effects of Different AFEX Treatments on Fluorescence Profile of Rice Straw 92 Figure 6.1 Replicate DRIFT spectra of B111 106 Figure 6.2 Effect of Delignification on the DRIFT Spectra of Rice Straw .......... ... 108 Figure 6.3 Effect of Various AFEX treatments on DRIFT spectra 110 Figure 6.4 Comparion of Spectra with different Sugar Yields at 48hr Enzynmtic Hydrolysis 111 8. 9. ABBREVIATIONS . AFEX — Ammonia Fiber Explosion CPFR - Continuous Plug Flow Reactor CrI - Crystallinity Index DRIFT Spectrometry — Diffuse Reflectance Infrared Fourier Transform Spectrometry DB- Dry Biomass FPA - Filter Paper Activity FPU - Filter Paper Units FTIR Spectrometry - Fourier Transform Infrared Spectrometry HPLC — High Pressure Liquid Chromatography 10. IR — Infrared ll. MLR — Multivariate Linear Regression 12. PCA - Principal Component Analysis 13. PCR - Principal Component Regression xi 1 INTRODUCTION The cattle industry represents a major portion of the United States economy. With such a large cattle industry in the United States, there is great demand for high quality livestock feeds at acceptable prices. Feeds utilized by the livestock producer are forages, grains, protein supplements, and vitamin and mineral supplements. Forages are vegetative portions of grasses or legume plants containing a high proportion of fiber and low energy. Fiber, or the cell wall, is composed of cellulose, hemicellulose and lignin. Humans cannot digest fiber but livestock animals can digest a part of the cellulose and hemicellulose portion of the fiber. Crop residues, an example of forages, are parts of the plants left over in the field after harvesting the primary crop. Grains, such. as barley, corn, rice and wheat, are low in fiber and high in energy. Generally, grains, protein supplements, minerals and vitamins are added to the forage as a source of energy and protein to meet the animal nutrient requirements. Cattle are classified ruminants because they regurgitate and chew the forage material to better digest the cellulose in the cell wall. Ruminants have a dense population of microorganisms in the reticulum and rumen, which produce enzymes capable of digesting cellulose and other substances resistant to digestion. The bacteria in the rumen degrade the cellulose within the plant cell walls into simple sugars such as glucose. Next, the microbes ferment the glucose into volatile fatty acids. The cow utilizes the end products of the microbial fermentation as a source of energy and for production of milk. Crop residues are not useful for human consumption because they contain high fiber, and thus low energy. Instead they can be used in fact as livestock feed. There is an abundant supply of crop residues in the United States that are not being effectively utilized. Agricultural statistics of 1988 have shown that over 400 million tons of crop residues were produced in the United States (Sundstol, 1988). The cost of soybean meal is approximately, $164-171 per ton while the cost of crop residues such as rice straw $ 22-35 per ton (Kansas City Weekly Feedstuff Review, Oct 13, 2000). Crop residues are economically beneficial to the livestock producer. Crop residues are in excess supply and are low in cost relative to other fiber sources such as alfalfa and therefore should be utilized more effectively as livestock feeds. Crop residues are a low quality feed because they lack many essential nutrients and cannot be adequately digested by the cow. Since grain is harvested from the plant after it has reached maturity, the crop residues contain high levels of lignin and are low in protein and digestible matter (Huber, 1981). These residues containing lignin are also known as lignocellulosic materials. If the low quality lignocellulosic materials could be converted to higher quality feeds, at acceptable costs, then feed could be provided for millions of animals from otherwise poorly utilized crop residues. The goal of the livestock producers is to feed cattle with the highest quality feeds at the lowest cost possible because a higher quality feed leads to increased beef production with greater profits for the livestock producer. The usual methods to improve the nutritional value of the crop residues are physical, chemical or physiochemical. There are other not so common pretreatment methods such as enzymatic hydrolysis of the biomass but enzymatic hydrolysis had not yet been improved in order to reduce the cost of the consumption of enzymes (Ogier et al., 1999). After the crop residues (lignocellulosics) are treated and processed as feed, they must compete both economically and nutritionally with conventional animal feeds to be utilized by the livestock producer. Most of the pretreatment methods improve feed quality by adding nutrients to the feed, by increasing the intake of feed by the animal or by improving the digestibility of the feed. The pretreatment methods that we are going to focus on are the ones, which increase the digestibility of the feed. 1.1 Treatments Physical Treatments The two common physical treatments of lignocellulosic materials are ball milling and two-roll milling. These physical treatments are expensive mainly due to the high energy and capital costs of grinding and pelleting the feeds (Holtzapple et al., 1991), 'which decrease the particle size and increase surface area for enzyme action (Abraham and Kurup, 1997). Furthermore, the chemical pretreatrnents (Sodium Hydroxide, Sodium Hydroxide-Acetic acid, chloroform, acid, sodium sulfite, peracetic acid, butanol, hydrogen peroxide-manganous salt, hydrochloric acid-zinc chloride, acetic acid—hydrogen peroxide) are much more effective on the hydrolysis of 1i grlocellulosics than physical pretreatments such as milling (Abraham and Kurup, 1997). Chemical Treatments Most chemical treatments involve hydrolytic and oxidative agents. Enzymatic hydrolysis studies of partial acid treated lignocellulosics show that 100% of the potential glucose content can be obtained after 24 hr enzymatic hydrolysis (Knappert et al., 1980). Peracetic acid has also been used for pretreatment of sugar cane bagasse (T eixeira et al., 1999) but before fermentation, the biomass had to be washed and neutralized to remove the acid. The weakness of using acids for pretreatment is, low recovery of the acid, formation of toxic products which inhibit fermentation of hydrolyzed sugars and degradation of biomass during pretreatment (Ogier et al., 1999). Lime treatment of the crop residues for 1-3 hr at high temperatures (85-135 °C) achieved a 93.6% biomass yield, 14% lignin solubilization but the problem was the recovery of lime (Chang et al., 1998). After 10 washings, about 86% of the added calcium was removed but soluble sugar is lost while recovering the calcium. Other treatments using solvents, organosolv and autohydrolysis are not very efficient because of their low recovery, low reactivity or because the biomass is easily degraded during treatment (Holtzapple et al., 1991). Chemical treatments using alkali (NaOH or NH3) increase the hydrolysis rate but again the problem with using sodium hydoxide is the low recovery of sodium hydroxide after treatment (Holtzapple et al., 1991). Currently in the US, the most widely utilized method of chemical treatment is ammoniation (Sundstol and Coxworth, 1984; Males, 1987; Mason et al., 1988). Anhydrous NH3, NH40H and urea are all sources of ammonia, which are added to the feed for pretreatment. Ammoniation improves the quality of the feed by improving digestibility of the feed (Males, 1987), increasing the intake by the animal (Morris and Mowat, 1980), and increasing the nitrogen content of the diet (Sundstol and Coxworth, 1984). Less protein needs to be supplemented to the feed due to the non-protein nitrogen from the ammonia. Only ruminants can use the non-protein nitrogen as a source of nitrogen for synthesis of proteins. Using ammonia for pretreatment has many benefits but ammonia does not react with the biomass as efficiently as the other chemical agents (Holtzapple et al., 1991). Physicochemical Pretreatments Physiochemica] reactions involve a chemical reaction but at the same time they also change the physical structure of the plant biomass. Although lignin is altered with chemical treatments, the lignocellulosic materials are chemically recalcitrant. Physicochemical treatments increase the accessible surface area and as a consequence, the lignocellulosic material is more susceptible to enzymatic hydrolysis (Holtzapple et al., 1991). A physiochemical treatment has the advantage of a physical treatment without the expense of high energy and capital costs. Two examples of physicochernical treatments are steam explosion and Ammonia Fiber Explosion (AFEX). Steam Explosion, a physiochemical process effectively disrupts the wood biomass and cleaves the glycosidic linkages in cellulose. The major disadvantage of using steam explosion is the loss of a large amount of hemicellulose sugars and a loss of 10% to 20% of the biomass during the treatment (Carrasco et al., 1994). The steam explosion process also operates at a very high temperature (235 °C) and pressure (32 atm), conditions that are detrimental to protein containing biomass (Holtzapple et al., 1991). A more recently developed ammoniation process, called AFEX, effectively treats non-woody biomass (Holtzapple et al., 1991) 1.2 Ammonia Fiber Explosion (AFEX) In the AFEX process, the biomass is soaked at high pressure (15 atm) and moderate temperatures in a pretreatment reactor. After completion of the pretreatment, a valve is opened at the bottom of the reactor through which the biomass is released into a flash tank. The sudden release in pressure causes the ammonia to flash and consequently causes the fibers of the biomass to explode. From the flash tank, the biomass is transferred to a dryer, which removes the residual ammonia. The amount of flash depends on the reaction temperature, pressure, moisture content and ammonia loading in the AFEX pretreatment reactor (Dale and Moreira, 1983). Only about 0.5 - 1% of ammonia is left in the biomass as chemically bound ammonium ions (Dale and Moreira, 1983). The residual ammonia will be used as a nitrogen source for the microbes. AFEX treatment has both the advantages of chemical treatment and physical treatment. Ammonia is inexpensive compared to other chemicals and it is volatile and therefore, easily recovered. Few if any fermentation inhibitors or sugar degradation products are produced during the AFEX treatment (Wang et al., 1998). The AFEX process has modest energy and capital costs, high reactivity and low degradation of biomass (Holtzapple et al., 1991). Mes-Hartree M. et al., (1988), found that AFEX treated biomass has a greater reactivity than steam exploded biomass. 1.3 Problem The feed stocks fed into the AFEX process are highly variable. The proportion of lignin, hemicellulose and cellulose varies among plant species, plant parts, and stage of growth. It is these differences in the structure of the cell wall, which cause variability in the digestion of lignocellulosic residues. The basic purpose of the AFEX treatment is to maximize the digestibility of the feed in the rumen of the animal. The parameters'for the AFEX treatment such as ammonia loading, water loading, temperature, pressure and reaction time, have to be chosen so that the digestibility of the feed is maximized. The current method (enzymatic hydrolysis) to determine digestibility of the AFEX treated materials is lengthy and expensive. We need to find a rapid, inexpensive and online technique to correlate the digestibility of the treated biomass to the initial conditions of the AFEX process or to some other factor of the rice straw. This technique will help us select the appropriate AFEX conditions. Before we decide on a technique, we need to first find out which features of the rice straw have the greatest impact on enzymatic hydrolysis. 2 LITERATURE REVIEW The plant cell wall is a complex biological structure containing a wide variety of molecules. A general overview of the plant cell wall is shown in Figure 2.1. / Primary Wall \ [ Secondary Wall \ Cell Lumen Li gnin k Low Concentration Gradient / High Figure 2.1 Schematic View of Plant Cell Structure Lignin is deposited in the primary wall and then gradually moves to the secondary cell wall always lagging behind polysaccharide deposition. Guaiacyl, p-hydroxyphenol and syringyl lignins are polyphenols derived respectively from, coumaryl, coniferyl, and sinapyl alcohols (Figure 2.2). DD ° Ho Ho HO / H coniferyl alcohol HO / H sinapyl alcohol Figure 2.2 Precursors of Guaiacyl, p-hydroxyphenol and syringyl Lignin Cell wall polysaccharides are composed of primarily of cellulose, glucans (heteroglucans, B—glucans), mannans(glucomannans) and xylans(glucuronarabinoxylans, heteroxylans 4- O-methyl-glucoroxylans). The glucans, mannans and xylans are grouped as one and called hemicellulose. The growth and development of the cell wall is divided into two stages. During primary growth, polysaccharides, proteins, phenolic acids, pectins, xylans and cellulose but no lignin are deposited in the cell wall causing the cell wall to elongate. The second. stage of development occurs at the end of the elongation process, when the cell wall begins to thicken. During the second stage, more cellulose than xylans is deposited (Bacic et al., 1988) and lignin depostition is initiated. Arabinoxylan ferulate esters of the primary wall cross-link the xylans to lignin (Iiyama et al., 1990). The p-coumaric acid is also incorporated into the cell wall with ether linkages to lignin polymer during the secondary wall development (Lam et al., 1990). Also, p-coumarate esters of 1i gnin are deposited into the secondary thickened wall (Jung and Vogel, 1992; Ralph et al., 1994). The polysaccharides, which are deposited, last are not all lignified but polysaccharides deposited in the primary wall region early on are very intensely lignified. Li gnin concentration is the greatest in the primary cell wall region. Here phenolic and non-phenolic acids can become linked to lignin and thus cause the li gnin to polymerize. Therefore, due to the steric hindrance of the enzymes by 1i gnin, digestibility of the mature cell wall is much less than for the immature cell wall due to lignification (Jung and Ralph, 1991). Cellulose is a major structural plant polysaccharide that consists of glucose molecules linked by 1,4-beta-glucosidic bonds. Since, the glucose molecules are rotated 180 degrees to each other, cellobiose is the repeating unit rather than glucose itself. mflllOH Cellobiose Figure 2.3 Repeating Unit of Cellulose Polymer The glucose molecules are in the chair configuration and therefore all the hydroxyl groups are in the equatorial position. Hydrogen bonding with other chains forms the rnicofibril of the cellulose. Numerous microfibrils form the macrofibrils. Since cellulose chains consist of 8,000 to 15,000 glucose residues, extensive hydrogen bonding can occur (Frey-Wyssling and Muhlethaler, 1963). The large number of 10 hydrogen bonds contributes to the mechanical strength of cellulose and its resistance to chemical degradation. The elementary fibrils are highly ordered in the center of the microfibril and thus form a rectangular crystalline core surrounded by a slightly disordered paracrystalline structure. A paracrystalline sheath covers the crystalline cellulose. The paracrystalline sheath in cotton contains mainly cellulose but in sources such as wood it also contains hemicelluloses and 1i gnin molecules. Some of the lignin and hemicellulose in the paracrystalline sheath maybe chemically bound to the cellulose. The lignin, hemicellulose and cellulose (lignocellulose) have a special three-dimensional relationship and through extensive cross-linking they comprise-the very rigid cell wall matrix of the plant. Lignin can be visualized filling the spaces between cellulose microfibrils, and between elementary fibrils in noncrystalline regions. The enzymatic hydrolysis of lignocellulosics is a heterogenenous reaction and is therefore, influenced by the structural features of the biomass such as crystallinity of cellulose, lignin content and surface area (Fan et al., 1980; Tsao et al., 1978). Crystalline cellulose refers to aggregates of cellulose polymers held tightly together by extensive hydrogen bonding and therefore water molecules are excluded from the crystalline inner structure. The inability of water to infiltrate the microfibril prevents hydration of the internal cellulose polymers of the microfibril, which in turn prevents cellulose hydrolysis by cellulolytic enzymes (Rowland, 1975). Lignin hinders the enzymes from coming in contact with the cellulose microfibrils and cross-linkage of lignin and cell wall polysaccharides such as cellulose can have a negative effect on hydrolysis rates (Jung and Allen, 1995). Also, polysaccharide saccarification may be limited due to the 11. hydrophobic nature of lignin (Van Soest, 1982). Surface area is defined as the accessible area for the enzyme. The accessible area for the enzyme is the external area of the lignocellulosic particle and the area of the pores (distance between cellulose microfibrils). Therefore, initial hydrolysis rates are dependent on the surface area of the lignocellulosic biomass (Grethlein and Converse, 1991). The different pretreatment methods are believed to increase the rate of enzymatic hydrolysis of the li gnocellulosic materials by modifying crystallinity of cellulose, 1i gnin content and surface area (Fan et al., 1981; Chang and Holtzapple, 2000). Physical pretreatrnents alter the crystallinity and increase“ the surface area of the substrate (Gharphuray et al., 1983). Crystallinity index of cellulose declined sharply with increasing ball-milling time but remained unchanged after 1hr (Gharphuray et al., 1983). The x-ray measurements of ball—milled cellulose by Fan et al. (1980) indicated a decrease in the crystallinity index. They found that the crystallinity index of ball-milled Solka Floc changed from 74.6 to 4.9 after 96 hours of milling. Ball milling not only reduces the particle size but also alters the crystalline structure of cellulose. Enzymatic hydrolysis studies of partial acid treated 1i gnocellulosics done by Diane Knappert et al. (1980) showed that acid treatment increased the rate of saccarification of lignocellulose due to removal of hemicellulose, reduction of the degree of polymerization and possibly change in the structure of the crystalline cellulose. Degree of polymerization is defined as the number of repeating cellobiose units of cellulose. Dilute acid treatment and autohydrolysis treatment of wood samples increased the pore volume (surface area) of the wood by removing hemicellulose (Grethlein and Converse, 1991), which in turn increased the hydrolysis of wood biomass. 12 Chemical treatments using alkali (N aOH or NH3) increase the hydrolysis rate of lignocellulosics by increasing their swelling capacity. The increase in swelling of the fibers is due to the saponification of esters of the 4-0-methyl glucuronic acid attached to xylan chains. Before treatment the esters act as crosslinks, limiting the swelling or distribution of polymer segments in water (Abraham and Kurup, 1997). Swelling of the fibers increases the surface area, which in turn increases the suseptibility of lignocellulose to enzymatic hydrolysis. Soaking of lignocellulosic biomass materials at high pressure in the AFEX pretreatment reactor causes decrystallization of cellulose, prehydrolysis of hemicellulose, and perhaps alteration of lignin structure (degradation of lignin polymers) (Holtzapple et al., 1991). The disruption of the fibers increases the surface area of the biomass, which in turn increases the enzymatic hydrolysis rate (Holtzapple et al., 1991). Cellulose decrystallization and increased surface area of the biomass, and perhaps the removal of some lignin significantly increase enzymatic digestibility of lignocellulosic biomass (Holtzapple et al., 1991). In the past, several investigators have examined the relationship between the hydrolysis rate and the important structural features (Baker, 1973; Cowling, 1975; Norkrans, 1950; Stone et al., 1969; Van Soest, 1969) but a recent analysis of the relationship was summarized by Chang and Holtzapple (2000). Some have suggested that physical features play an important role in biomass digestibility. Some of the conclusions were: a) degree of polymerization has an inverse relationship to hydrolysis because it lowers the accessibility of cellulose to solvents (Puri, 1984), b) an increase in pore volume increases initial saccharification rate (Grethlein, 1985), c) surface area (pore 13 volume and particle size) has a positive influence on hydrolysis rate (Puri, 1984; Grethlein, 1985). Many studies found an inverse correlation between lignin content and digestibility (Sullivan, 1959; Van Soest, 1969; Feist et al., 1970; Baker, 1973; Fan et al., 1981; Gharpuray et al., 1983; Kong et al., 1992; Thompson et al., 1992; Koullas et al., 1992; Vinzant et al., 1997). Others found an inverse correlation between crystallinity of cellulose and digestibility (Norkrans, 1950; Walseth, 1952; Sasaki et al., 1979; Fan et al., 1980; Fan et al., 1981; Gharpuray et al., 1983; Bertran and Dale, 1985; Weimer and Weston, 1985; Rivers and Emert, 1988; Thompson et al., 1992; Koullas et al., 1992). According to Chang and Holtzapple (2000), these studies had many limitations due to the small sample size and the cross effects among the various structural features. They contend that lignin content and the crystallinity index have the greatest impact on biomass digestibility. Chang and Holtzapple (2000) studied lignin content, crystallinity index and acetyl content as three independent variables. About hundred and forty seven of hybrid poplar, switchgrass and bagasse samples were selectively delignified using peracetic acid, deacetylated using potassium hydroxide and decrystallized using the ball milling technique. Lignin content, moisture and carbohydrate concentration was established using the NREL standard; acetyl groups were determined using the transesteliflcation procedure (Browning, 1976) while crystallinity index was measured by the method developed by Sega] et al., 1959. Sugar values from enzymatic hydrolysis of the delignified, deacetylated and decrystallized samples were measured. Correlation studies then tried to find a relationship between enzymatic hydrolysis and the three independent variables. The studies led to the following conlusions (Chang and Holtzapple, 2000): 14 . Low lignin is sufficient to obtain high digestibility regardless of Cd or acetyl content. Low crystallinity is sufficient to obtain high digestibility regardless of lignin or acetyl content. . Hydrolysis rate is not dependent on diffusive mass transfer but rather on the quantity of adsorbed enzymes and their effectiveness (effectiveness is a function of Cd). . Samples with low lignin and Cd can adsorb more enzymes and therefore can be rapidly digested. . Samples with low lignin but high CrI can adsorb enzymes but the enzymes are not as effective due to the resistant crystalline core of cellulose. . Samples with high lignin but low CrI can adsorb lesser amounts of enzymes but the adsorbed enzymes are very effective and almost total digestion can occur within 3 days. . For samples with high 1i gnin and high CrI, the enzymes are blocked by lignin and also the ezymes are not very effective because of the high degree of crystallinity and the digestion rate is poor. . Lignin content, acetyl content and Cd completely determine enzymatic digestibility but deacetylation had a 15 greater effect on hemicellulose digestibility than cellulose digestibility. Therefore, they concluded that lignin content and Cd have the greatest impact on the digestibility of biomass. 2.1 Hypothesis Our hypothesis was based on Chang and Holtzapple’s (2000) research. We postulate that the conditions in the pretreatment reactor affect the composition of the AFEX treated biomass. Since prior research has already shown that the AFEX treated biomass has greater extent of hydrolysis than untreated biomass (Holtzapple et al., 1992) we can state that the AFEX treatment is affecting some feature of the biomass. We predict that the Cd and the lignin content of the AFEX treated biomass will vary with variation of treatment conditions. Also, we predict that the treatments may have a greater impact on one of these features and therefore, that feature may determine the digestibility of the treated biomass. Therefore, to optimize the process, we need to study the Cd and lignin content of the AFEX treated biomass. Since, the lignin content and Cd have the greatest impact on hydrolysis of lignocellulose, analytical techniques that can measure these two factors need to be chosen. Also, the selected technique has to be rapid, inexpensive and online. 2.2 Technique to Determine Crystallinity Index (CrI) Since we could not find a technique to measure both Cd and ii gnin content together, we had to use two separate techniques. Techniques such as inelastic neutron scattering (Muller et al., 2000), thick layer wicking (Dourado et al., 1998), staining (Y u et al., 1998), solid-state NMR (EK et al., 1994), electron-diffraction (Paralikar et al., 16 1997) and spin echo NMR (EK et al., 1994); have been tried to study the crystalline structure of cellulose. These techniques studied cellulose beyond what was necessary for our experiments and also, they were expensive. X-ray diffraction (Segal et al., 1959) techniques were used to determine the crystallinity index of cellulose. Therefore, we chose x-ray diffraction as the technique to determine the CrI of AFEX treated biomass. The Cd is a measure of the relative amounts of the crystalline and amorphous regions in cellulose. The formula for detennining the CrI proposed by Sega] et al. (1959) was: CrI = ((IOOZ'Iam)/Iam) X 100 1002 is the intensity of the peak at 20:22.8° Iam peak is the intensity of the 20=18° 1002 represents the crystalline fraction of the cellulose while the Iam represents the amorphous cellulose. The intensities of the two peaks were measured above an approximate baseline representing background intensities. Cellulose is the largest contributor to the measured amount of crystallinity, however, other components such as hemicellulose and lignin, may contribute slightly (Gharpuray et al., 1983). 2.3 Technique to determine Lignin Content Due to the need for rapid analysis, the lignin cannot be extracted and the samples should be analyzed as a solid. Therefore a spectrophotometric analysis of lignin in the solid plant material would be the optimum technique to try. Chemolytic techniques with calorimetric analysis (Kogel-Knabner I, 2000), FTIR (Silva et al., 1999; Rodrigues et al., 1998), solid-state C-13 NMR (Love et al., 1992), fluorescence (Olmstead and Gray, 1997; Billa et al., 1999), DRIFT (Backa and Brolin, 1991; Pandey 1999) were all used 17 for the determination of lignin content in pulp or in other lignocellulosic materials. Among the various techniques, Fluorescence and Diffuse Infrared Fourier Transform (DRIFT) were the optimal techniques to analyze AFEX treated biomass materials. Both the techniques analyze the biomass as a solid material and fluorescence spectrophotometers and IR machines are readily available in most labs. Research done by Billa et al. (1998) gave a good correlation between the fluorescence data and the extracted lignin content of wheat straw. According to them, fluourescence spectroscopy is a cheap, rapid, specific and non—destructive method for determining lignin content. DRIFT was used to determine the lignin and carbohydrate content of wood pulp (Backa and Brolin, 1991) and most recently, DRIFT spectroscopy was used for the quantitative analysis of major wood components concentrations at the 99% confidence interval representing r2 values higher than 0.86. DRIFT is a simple technique to monitor structural changes during chemical or physical processing (Owen and Thomas, 1989; Michell, 1994) and it requires small sample sizes and short analysis times. Fluorescence and DRIFT techniques are rapid, inexpensive, on-line, non- destructive and non-intrusive. They have an advantage over other conventional chemical methods, which are time consuming and also result in the degradation of natural polymers. We are assuming that lignin content can be predicted using these techniques and since, lignin content has a dramatic impact on the digestibility of biomass, these two techniques should be able to predict the lignin content in rice straw and correlate with the digestibility of biomass. Therefore, these two techniques have been chosen to predict the digestibility of AFEX treated biomass by determining the content of lignin. 18 2.4 Problem Solving Approach 1. Determine the sugar yield of the AFEX treated biomass from in-vitro enzymatic hydrolysis studies using cellulolytic enzymes. Analyze the samples using X-ray diffraction, fluorescence and DRIFT techniques. Correlate the concentration or percent yield of sugars to the analytical data using multi-linear regression or principal component analysis. Validate the models using the R2 values (R2 value closest to 1 is preferred). 19 3 HYDROLYSIS 3.1. Raw Materials The rice straw samples were treated at Texas A&M University using the Ammonia Fiber Explosion (AFEX) process. The untreated rice straw was ground to pass a 2mm screen prior to AFEX treatment. The rice straw samples were pretreated in a batch reactor with high-pressure liquid ammonia. Next, the pretreated biomass was released into a flash tank through the opening of a valve at the bottom of the reactor. The instantaneous drop in pressure causes the liquid ammonia to flash to vapor, which in turn causes the explosive decompression of the rice straw. From the flash tank the AFEX treated samples were sent to a drier to remove residual ammonia. The extent of the AFEX explosion depends on the pretreatment process parameters such as temperature, reaction time, water loading and ammonia loading. Currently the AFEX process is undergoing optimization. Therefore, the rice straw samples were treated at different process parameters to study the effect of treatment conditions on digestibility. The various treatment conditions of the AFEX process are specified in Table 3.1. 20 Table 3.1 AFEX Process Parameters for Treated Rice Straw Treatment Treatment Moisture Treatment Treatment Ammonia ID (%)* Time (min) Temperature Ratio (**) ( °C ) B103 60 10 90 05:10 B1 04 60 10 80 2.0:1.0 B105 60 10 90 0.75:1 0 B108 60 10 90 1 .0:1.0 B1 1 20 5 80 1 .0:1.0 B1 1 1 60 10 80 1 .0:1.0 B1 14 20 5 80 1 .5:1.0 B1 20 20 5 90 1 .5:1.0 B123 20 10 80 1 .5:1.0 B126 20 10 80 2.0:1.0 B129 20 5 80 2021.0 B135 20 10 90 15:10 B1 38 20 10 90 2.0210 B14 20 10 80 05:10 B141 40 5 80 1 .5:1.0 B144 40 5 80 20:10 B1 47 40 5 90 15:10 B1 50 40 5 90 2021.0 B153 40 10 80 1 .5:1.0 B1 56 40 10 90 1 .5:1.0 B1 59 40 10 90 20:10 B162 40 10 80 2.0:1.0 B1 65 60 5 80 15:10 B1 68 60 5 80 20:10 817 20 10 80 0.7 :10 B177 60 5 90 1 .5:1.0 B180 60 5 90 20:10 81 86 60 10 90 20:10 B2 20 5 0521.0 * Treatment Moisture ((Kg water/Kg dry straw)x100) **Ammonia Ratio (lb ammonia/lb dry straw) 21 Table 3.1 (Cont’d) Treatment Treatment Treatment ID Moisture 3:133:13 Temperature Ammonia (%) * ( °C ) Ratio (**) 820 20 10 90 05:10 B24 20 10 90 0.75:1.0 B28 20 10 90 10:10 B32 20 5 90 01:10 B35 20 ' 5 90 0.75:1.0 B38 20 5 90 10:10 B41 20 10 80 10:10 B45 40 5 80 0.75:1 .0 B5 40 5 05:10 851 40 5 90 05:10 854 40 5 90 0.75:1.0 857 40 10 80 0.5:1 .0 B60 40 10 80 0.75:1.0 B63 40 5 90 10:10 866 40 10 80 10:10 869 40 10 90 0.5:1 .0 872 40 10 90 0.75:1.0 B75 40 10 90 0.1 :1 .0 B78 60 5 80 05:10 88 20 5 80 0.75:1.0 882 60 5 80 10:10 884 60 5 80 10:10 887 60 5 90 0. 5: 1 890 60 5 90 0.75: 1 .0 893 60 5 90 1.0: 1.0 0172 60 10 80 1 .:5 1.0 C184 60 10 90 1 .:5 1.0 C49 40 5 80 1 .:01 .0 * Treatment Moisture ((Kg water/Kg dry straw)x100) **Ammonia Ratio (lb ammonia/lb dry straw) 22 Each sample of rice straw was treated at the AFEX conditions specified. A limited set of pretreatment conditions was considered. Temperature for the treatment ranged from 80 to 90 °C while moisture varied from 20, 40, and 60 percent at a reaction time of 5 or 10 min. The ammonia ratio varied from a ratio of 0.1 to a final ratio of 2.0. Combinations of the above mentioned parameters were used to treat the rice straw. Prior research determined that this range of conditions exhibited the full range of the AFEX response to treatment; from very little response to essentially complete response. The AFEX rice straw samples were powdered for hydrolysis studies. Some of the AFEX treated rice straw samples already contained finely crushed straw probably due to shipping. Other samples were powdered using a mortor and pestle. Untreated rice straw was used as a control and it was powdered with a mortor and pestle also. All the samples were then sieved through an Opening of 0.106 mm (ISO-mesh) to obtain a homogeneous powder. 23 3.1 Experimental 3.1.1 Delignification of Rice Straw Untreated rice straw was delignified according to the method of Van Soest and Wine (1968) except wedid not make acid detergent fibers first. Our goal was to just delignify rice straw and the permanganate oxidation with the preceeding steps was enough to attain delignification. 3.1.2 Determining moisture content of rice straw The dry weight of the samples was determined by drying the rice straw overnight in a vacuum oven adjusted at 80 °C. The empty aluminum weigh boats were weighed and then tared using an analytical balance sensitive to 0.0001 g. Then about 0.2 g of the . lice straw was placed in the aluminum weigh boats and the samples were weighed. The samples were dried overnight at 80°C. The samples were weighed prior to drying and immediately after drying. Subtraction of the weight of the aluminum weigh boats gave us the actual weight of the sample. Moisture content was determined as follows: Moisture Content (%) = 100 * ((Mass of wet weight of sample) - (Mass of dry weight of sample» (3.1) ( Mass of wet weight of sample) Since a large number of samples were being dried at the same time, the samples were placed in a container containing anhydrous calcium sulfate after being dried overnight. Thus the rice straw samples did not regain any moisture prior to reweighing. 3.1.3 High Pressure Liquid Chromatography (HPLC) A I-IP-X-87X Ion Exclusion column (300 X 7.8 mm) was used to separate the sugars. The mobile phase consisted of 0.01 N H2804 set at a rate of 0.4 mL/min and a 24 temperature of 60°C. Standards for the HPLC were made from a 10 mg/mL stock solution. The stock solution was kept at 22 °C to avoid concentration gradients in the solution. The concentrations of the standards were: 2 g/L, 1 g/L, 0.5 g/L and 0.25 g/L. Dilutions were done meticulously using sterile methods. To make sure that the samples were not contaminated, the standards were analyzed on the HPLC. The standards gave us predicted concentrations and an r2 value of 0.99. 3.1.4 DIONEX The procedure for the DIONEX system was set to the specifications given in DIONEX Technical Note 20 (Analysis of Carbohydrate by High Performance Anion Exchange Chromatography with Pulsed Amperometric Detection). A CarboPac PA-lO (4 X 250mm) analytical column was used to analyze the sugars. 3.1.5 Acid Hydrolysis The AFEX treated rice straw samples were dried overnight at 80 °C. Approximately, 0.1g of the dry rice straw was hydrolyzed in 1 mL of a 72% sulfuric acid solution for 1 hr in a 30°C water bath. The rice straw was hydrolyzed in 13 x 100 inch test tubes. The mixture was stirred every 10 min with long glass stining rods. After 1 hr the hydrolyzed mixture was quantitatively diluted by adding 28 mL of distilled water and then autoclaved for 45 min in 125mL Erlenmeyer flasks covered with aluminum foil. Upon cooling the mixture was filtered through Whatrnan No. 1 filter papers using a celite funnel and a vacuum flask. Finally, the filtered liquid sample was further diluted to a final volume of 100 mL. The samples were filtered immediately using a 0.45 pm Gelman filters into three HPLC vials and frozen straight away at (~20 °C). Although only 3 mL of the acid hydrolyzed was filtered, the rest of the sample was also frozen for 25 future need. Prior to HPLC analysis, the frozen samples were thawed and well mixed by turning the vials upside down. The samples were analyzed for sugars using high- pressure liquid chromatography (HPLC). 3.1.6 Enzymatic Hydrolysis Approximately 0.5 g of rice straw, according to dry weight, was added to a final volume of 10 mL 0.1 M citrate buffer (pH 4.8) containing 20 ppm of sodium azide. All solutions and the biomass were assumed to have a specific gravity of 1.0 g/L. The amount of rice straw according to dry weight, added to each test tube was calculated as follows: (0.5 g dry weight) (32) Rice straw aceordin to d wei ht = g ry g (g) (1 - (Percent moisture content/10 0)) 1 The percent moisture content was determined as specified in Equation 3.1. The cellulase enzyme (Celluclast, Novo Nordisk) activity was 71 FPU/mL. Activity of the cellulase enzyme was determined by the method of Mandel (1976) and Miller (1959) using Whatrnan No. 1 filter paper as a substrate and was expressed as filter paper activity (FPA) in terms of filter paper units (FPU). Cellulase loading was 5 IU/g dry biomass while Cellobiase (B-Glucosidase) loading was 30 IU/g dry biomass. Three times the volume of the Celluclast was used for cellobiase loading. The activity of the cellobiase was 150 FPU/mL (B—Glucosidase, Novo Nordisk). The rice straw was hydrolyzed at 50 °C for 48 hrs in a ISO-rpm water bath shaker in 13 x 100 inch test tubes. The contents of the test tube were warmed to 50°C before the addition of enzymes. The test tubes were covered with parafilm during hydrolysis. The shaking of the test tubes allowed the enzymes and the rice straw to mix properly. The 26 hydrolyzed samples were well mixed before taking 1mL of sample at time points 0 hr, 3 hr, 24 hr and 48hr. Mixing before removing the sample at a certain time point allowed the concentration of the enzyme to be uniform. The samples were transferred using a glass pipet attached to an eppendorf tip. The bottom part of both the glass pipet and eppendorf tips were cut to allow solids as well as liquid to be withdrawn into the orifice. The samples were frozen at (—20°C) until they were analyzed. Prior to analysis the samples were thawed and then centrifuged for 5 min. The supernatant was diluted 100 fold and filtered through 0.45 pm Gelman HPLC certified filters prior to HPLC analysis. 3.1.7 Statistical Modeling Multivariate linear regression was done between the initial conditions and the concentration of each sugar. The analysis of variance (ANOVA) for each model gave 1'2 values that were used to estimate the quality and validity of the models. 27 3.2 Results and Discussion In vitro digestion studies were done for all the AFEX treated lice straw samples using cellulase and cellobiase enzymes. These studies helped us understand the effect of various AFEX treatment conditions on the concentration and percent yield of sugars. Prior to hydrolysis, factors affecting the enzymatic digestibility were considered. 3.2.1 Factors affecting Enzymatic Hydrolysis Activity of Enzyme One factor affecting the enzymatic hydrolysis is the pH of the rice straw. The pH of the rice straw affects the activity of the Celluclast (cellulase) enzymes and thus affects the ability of the Celluclast to digest the rice straw. The Celluclast is most active at a pH of 4.8. Since the AFEX rice straw samples were treated at various ammonia ratios, the pH of the rice straw was checked. The pH of the various straw samples was measured by soaking the rice straw in 10 mL of 0.5 M sodium citrate buffer, pH 4.8. Among the rice straw samples treated with different ammonia ratios, the pH of the soaked sample varied from pH 4.98-5.03. The pH remained constant even after the samples were soaked for 48 hr in 10 mL of 0.5 M sodium citrate buffer, pH 4.8. The pH of the rice straw did not vary significantly among the various samples. Since the pH does not vary much, we can safely assume that the activity of the Celluclast is not affected by the treatments with various ammonia ratios tested. Washing of the rice straw before hydrolysis was not necessary. Washing should be considered if there is a considerable change in the pH of the rice straw after AFEX treatment. Also, the stability of the Celluclast and the B—glucosidase enzymes was tested. If the activity of the enzymes changed over period of time, then the amount of enzyme 28 added to 05ng biomass would vary with time. Initially the activity of the Celluclast decreased but over a period of time it reached a constant activity level of 71.04 FPU/mL. All the enzymatic hydrolysis studies were performed when the activity of the Celluclast stabilized to 71.04FPU/mL. The activity of the B-glucosidase remained at a constant activity level of 150TU/mL. Moisture Content The percentmoisture content of the AFEX treated rice straw was measured to determine the dry weight of the sample. Table 3.2 gives the percent moisture content of the AFEX treated rice straw samples determined by drying the samples at 80 °C overnight. 29 Table 3.2 Moisture Content of the AFEX Treated Rice Straw Treatrnen Percent 10 ngiilergt Treatment Tli‘neapzraetsi'e Ammonia:i Moisture . , time (min) , ratio content W (c) (”) (m) 8103 60 10 90 05:10 5.37 81 04 60 10 80 20:10 4.49 8105 60 10 90 0.75:1.0 5.80 8108 60 10 90 10:10 5.70 811 20 5 80 10:10 4.31 8111 60 10 80 1.0210 5.07 B114 20 5 80 1.5:1.0 5.28 8120 20 5 90 15:10 4.51 8123 20 10 80 1.5210 4.69 8126 20 10 80 20:10 4.60 8129 20 5 80 20:10 5.96 8135 20 10 90 1.5:1 .0 5.53 8138 20 10 90 2.0:1 .0 3.90 814 20 10 80 0521.0 5.27 8141 40 5 80 15:10 5.14 8144 40 5 80 20:10 4.97 8147 40 5 90 15:10 4.38 8150 40 5 90 2.0:1 .0 5.22 8153 40 10 80 15:10 5.33 8156 40 10 90 15:10 5.23 8159 40 10 90 2.0:1 0 5.26 8162 40 10 80 2.0:1 .0 5.20 8165 60 5 80 15:10 5.53 8168 60 5 80 2.0:1 .0 4.54 817 20 10 80 0.75:1.0 4.19 8177 60 5 90 1.5:1.0 4.65 8180 60 5 90 2.0:1 .0 5.08 * Treatment Moisture ((Kg water/ Kg dry straw)x100) ** Ammonia Ratio (lb ammonia! lb dry straw) *** Percent Moisture Content ((grams HzO/ grams total sample) x100) 3O Table 3.2 (Cont’d) Treatment Treatment Treatment name." ”5”“ ID Moisture Time Temperatu “@2333 “23:33: (%): (min) (°C) (a, (..., 3186 60 10 90 2021.0 5.08 32 20 5 0521.0 4.03 320 20 10 90 0.521.0 3.84 324 20 10 90 0.7521.0 4.12 328 20 10 90 1.021.0 4.67 332 20 5 90 0.1 21.0 3.25 335 20 5 90 0.7521 .0 3.75 341 20 10 80 1.021.0 4.51 345 40 5 80 0.7521 .0 5.74 35 40 5 0.521 .0 5.28 351 40 5 90 0.521.0 4.12 354 40 5 90 0.7521 .0 5.03 357 40 10 80 0.521.0 4.63 360 40 10 80 0.7521.0 4.48 363 40 5 90 1.021.0 5.72 366 40 10 80 1 .021.0 4.92 369 40 10 90 0.521.0 4.80 372 40 10 90 0.7521 .0 6.45 375 40 10 90 0.121.0 5.45 378 60 5 80 0.521.0 5.36 * Treatment Moisture ((Kg ammonia/ Kg dry straw) x100) ** Ammonia Ratio (lb ammonia/ lb dry straw) *** Percent Moisture content ((grams H20! grams total sample)x100) 31 As shown in Table 3.2 the percent moisture content of the treated straws varied from 3.2 to 6.4%. Sample B32 with the smallest percent moisture content was treated at a temperature of 90°C for 5 min at 20% moisture and an ammonia ratio of 0.75. Sample B72 with the largest percent moisture content was treated at a temperature of 90 °C for 10 min with 40 % moisture content and an ammonia ratio of 0.75. Untreated rice straw had moisture content of 5.45%. Replicating the percent moisture content of the samples gave a 2% error. Moisture content did not vary a lot and therefore the percent moisture content did not affect the actual weight of the rice straw by’ much. The conventional method for determining dry weight is by drying the sample at 105 °C for 8 hrs. We checked to see how much the moisture content varied, by drying Sample 3165 at 105 °C and at 80 °C for 8 hrs. Acid hydrolysis of 0.5 g of samples from both drying methods gave an approximate dextrose concentration of 0.32 g/L. Also the increase in the moisture content of the sample was less than 2%, which was within the error. Therefore, drying the samples overnight at 80 °C did not affect the dry weight of sample by much. 3.2.2 Composition of rice straw During enzymatic hydrolysis, the cellulase breaks down the cellulose into [3- glucose units otherwise known as dextrose. The cellulase mixture is comprised of endoglucanases, cellobiases and exoglucanases. Endoglucanases convert noncrystalline cellulose to cellooligosaccharides by attacking the free cellulose chains. The cellobiases convert the low number cellooligosaccharides and cellobiose to B-glucose. The exoglucanases remove cellobiose units from the non-reducing ends of cellulose chains and may act synergistically with endoglucanases to convert native crystalline forms of 32 cellulose to cellooligosaccharides. Hemicellulose, on the other hand, has a backbone of five carbon sugars such as xylose with lesser amounts of arabinose, galactose, glucose and uronic acids branching off of it. The cellulase also digests the hemicellulase and other polysaccharides but not as efficiently as it digests cellulose. Xylose Figure 3.1 Sugars from Hydrolysis Studies Sugars obtained from acid hydrolysis of the AFEX treated rice straw determined the composition of the rice straw. Some of the acid hydrolyzed rice straw samples were analyzed on the DIONEX systems to determine the composition of total sugar. The DIONEX system is sensitive for detection of arabinose, galactose, glucose, xylose and mannose. Only the concentrations of glucose and xylose were significant from analysis on the DIONEX system. Negligible amounts (<1%) of other sugars were present in the sample analysis report Therefore, total sugar within the rice straw is composed of primarily xylose and dextrose. Similar results were seen in the hydrolysis of wood pulps (Chang and Holtzapple, 2000). Analysis of Sugars Unlike the DIONEX system, the HPLC is not sensitiVe to the detection of arabinose, galactose and mannose. HPLC analysis does not fully separate the sugars from one another and many of the sugars elute along with xylose as one peak. Since 33 only xylose and dextrose are the significant products of enzymatic hydrolysis, the hydrolyzed samples were analyzed on the HPLC. Glucose and xylose eluted as two separate peaks by the HPLC analysis method. The concentrations of dextrose and xylose at time 3 hr, 24 hr and 48 hr for the hydrolyzed samples (0.5 g AFEX-treated rice straw) are respectively, given in Tables 8.1, 8.2 and 8.3 in the Appendix. All the AFEX treated samples showed an increase in digestibility as compared to untreated rice straw. Reproducibility Triplicate rice straw samples were used to assure reproducibility. Two samples were hydrolyzed at the same time in the same water bath while another sample was hydrolyzed at a later time in another water bath. This was to ensure that the errors in the hydrolysis were independent of the conditions of the water bath and the solutions added to the hydrolytic mixture. The reproducibility error from enzymatic digestion at 348 hr for the concentration of dextrose was :30 %. The error for the concentration of xylose was 1:25 %. An error of i10% was seen in acid digestion studies. Therefore, some of the error can be attributed to the fact that not every strand of rice straw has the same composition but most of the error can also be blamed on the AFEX process. In this batch process, the temperature, ammonia level and moisture have not been fully regulated uniformly in the container. These various treatment conditions along the reactor could be the main cause of error in the measurement of the concentration of sugars." The amount of sugar from 0.5 g of dry biomass is determined by multiplying the concentration of sugar from the HPLC analysis by 100. Three hundred microliters were removed from the 1mL sample taken at the specific time point and diluted to a final 34 volume of 3mL. Therefore, the dilution factor is 10 fold and is accounted for in the equation. Amount of sugar from enzymatic hydrolysis of 0.5 g dry biomass (mg) (3.3) 3mL = (concentration of sugar from HPLC analySiS)x[ 0 3 . m )xlOmL Table 8.1 gives the concentration of sugars from the hydrolysis of 0.5 g dry biomass of the AFEX treated rice straw at 3 hr. The highest concentrations of digested sugars were seen for sample 8114 and 3138. Both samples had a dextrose concentration of 0.29g/L (29mgdextrose/O.5g DB) and both were treated at 20% moisture along with an ammonia ratio of 1.5-2.0. Among the treated samples, the lowest concentration of sugar was seen in sample B38 that gave a dextrose concentration of 0.06 g/L (6 mg/0.5g DB) and a concentration of 0.04 g/L (4mg/ 0.5g DB) for xylose. Sample B38 was treated at 20% moisture, 90°C for 5 nrin with a 1.0 ammonia ratio. Sample B38 had a lower concentration of dextrose than the untreated rice straw. Untreated rice straw had a dextrose concentration of 0.1 g/L (10mg/O.5 g DB) and a xylose concentration of 0.027g/L (2.7mg/O.5g DB). No conclusion can be drawn on how treatment conditions affect the initial hydrolysis rates of the biomass, with respect to sample B38, 8114 and 3138. At 24hr(Table 8.2) the samples did not retain the same rate of digestion. The highest concentration of 0.64g/L dextrose (64mg/0.5g DB) and 0.28g/L xylose (28mg! 0.5g DB) was seen with sample 3126. The concentration of sugar for 3126 tripled from the 3 hr time point. Sample 8126 was processed at 20% moisture for 10 min with an ammonia ratio of 2 and a temperature of 80°C. Samples 3180 and B87 were digested the 35 least. Both the samples were treated at 60% moisture, 90°C for 5min with varying ammonia ratios (2, 0.5). Untreated rice straw had a dextrose concentration of 0.17g/L (17mg/0.5g DB) and a xylose concentration of 0.04 g/L (4 mg/0.5g DB) at the 24 hr time point. After 48 hrs (Table 8.3), untreated rice straw peaked at a dextrose and xylose concentration of 0.22 g/L (22mg/0.5 g DB) and 0.06 g/L (6mg! 0.5 g DB), respectively. Minimal sugar concentrations of 0.32 g/L (32mg! 0.5 g DB) dextrose and 0.11 g/L (11mg! 0.5 g DB) xylose were obtained for sample 335. The greatest rate of digestion was seen with samples 3147 and 3150 (0.73g/L dextrose, 0.36g/L xylose). They were both treated at 40% moisture, 90 °C for 5 min and a high ammonia ratio. Comparing samples 335, 3147 and 3150 we can conclude that; moderate moisture (40%), high ammonia ratio (1.5-2) are the optimal process conditions to obtain maximal hydrolysis of rice straw. 36 Trends in Sugar Concentration Trends in the rate of hydrolysis for some of the treated samples were plotted. Concentration of dextrose and xylose versus hydrolysis time are shown respectively, in Figure 3.2 and 3.3. 37 Trends in Concentration of Dextrose P on , - . . , . . . - f . . . . , . . . . , j l i +8150 ‘i o 7 + Untreated Rice Straw .1, f + 8114 1’ l + Delignified Rice Straw 1 A '. —-+— 3180 j 510-6 f 7 2‘3 v i a i i g 0.5 I; j; x L .' 8 i C i :z .2 i l. P . Iii 0 3 l j’ h F .1 o- .. 5 ‘ 3 o 0 2 - c .. o I o ' ‘ .4 0 1 f/ j l o . n . 1 l . l . . 1 1 i n A l 0 1 O 40 50 20 30 Hydrolysis time (hr) Figure 3.2 Trends in the Concentration of Dextrose B150 (40% Moisture, Ammonia Ratio 2.0, 90°C, 5min) 8114 (20% Moisture, Ammonia Ratio 1.5, 80°C, 5min) B180 (60% Moisture, Ammonia Ratio 2, 90°C, 5min) 38 0.5 .o 9 .° N 03 h Concentration of Xylose (91L) .° .5 Trends in Concentration of Xylose r ‘7 T f I I T T fl ~ . Y i T T I r T +8150 +Untreated Rice Straw + 81 14 +Deligni1ied Rice Stra -—'-—8180 1 0 20 30 40 Hydrolysis Time (hr) Figure 3.3 Trends in the Concentration of Xylose B150 (40% Moisture, Ammonia Ratio 2, 90°C, 5min) B114 (20% Moisture, Ammonia Ratio 1.5, 80°C, 5min) B180 (60% Moisture, Ammonia Ratio 2, 90°C, 5min) 39 50 As seen in Figures 3.2 and 3.3, some of the AFEX treatments can significantly affect the hydrolysis rate of lignocellulosic biomass. The same trend (order of sample with increasing digestibility) is seen for both digestion of xylose and dextrose. In the overall trend, untreated rice straw was digested the least while sample 3150 had the highest rate of hydrolysis. Untreated rice straw seems to have reached maximum digestion within 24 hr as seen by the almost flat line from 24 to 48 hr while the AFEX treated samples do not seem to have reached maximum digestibility even after 48 hr enzymatic hydrolysis. Also, as seen in the plot, initial (0-3hr) hydrolysis rates are the greatest for all the samples. Delignification had a definite impact on dextrose concentration but not on xylose concentration. The concentration of dextrose improved significantly while the concentration of xylose improved only slightly. Removal of the lignin barrier improves the digestibility of cellulose because the enzymes can now adsorb more easily to the substrate. Lignin has a slight impact on the digestibility of hemicellulose because of the removal of the lignin barrier but it does not have such an impact on hemicellulose as it does on cellulose. During the pretreatment process, some sugars were released. Zero time points determined the concentration of sugar (dextrose and xylose) released by the rice straw during pretreatment. These sugar concentrations varied from 0.05 to 0.02 g/L (2mg / 0.5g DB) of dextrose and from 0.05 to 0 g/L (5 ~0mg l0.5g DB) of xylose. Both sample 3150 and 3114 with the higher final hydrolysis rates had a greater concentration of sugar at time zero while sample B180 and unheated rice straw had a minimal concentration of 0.02 g/L dextrose (2mg/0.5gDB) and no concentration of xylse at time zero. Delignified rice straw had no sugars for time 0hr. Probably the washing of the fibers removed all the 40 sugars. The amounts of sugar released were at random, minimal and did not correlate with treatment conditions. Analysis of Percent Yield of Sugar Percent yield of xylose allows us to determine the approximate amount of hemicellulose digested and the percent yield of dextrose allows us to determine the amount of cellulose digested within the rice straw. Hemicellulose sometimes does contain glucose but it is primarily composed of repeating xylose units especially in our case, because we found minimal amounts of other sugars present. Also, the cellulase enzymes do not digest the hemicellulose as efficiently as the cellulose. The percent yield of dextrose and xylose at 3 hr, 24hr and 48 hr are given in Tables 8.4, 8.5, 8.6 in the Appendix. Percent yield of sugar is determined by using Equations 3.3, 3.4 and 3.5. Amount of Sugar from Acid Hydrolysis of 0.1 g Dry Biomass (mg) = (3 4) C oncentrat ion of Sugar from HPLC anlaysis x 100 (Amount of sugar from enzymatic digestion ) Percent Yield of Sugar(%) = * 100 (3.5) (Amount of sugar from acid digestion * 5) The amount of sugar from the enzymatic digestion was determined using Equation 3.3 while the total amount of sugar was determined using Equation 3.4. The concentration of sugar from acid digestion is multiplied by a value of 5, because only 0.1 g of rice straw was digested with acid while 0.5 g of rice straw was digested during enzymatic hydrolysis. The amount of sugar from acid hydrolysis is the potential amount of sugar in the rice straw. The small hydrogen ions can cleave the acetal bonds of the sugars while the large enzyme is hindered from hydrolyzing the substrate due to its size and 41 catalyztically active agents (Philipp et al., 1979). Therefore, the acid can hydrolyze all the bonds of the sugar polymers in the rice straw. The highest percent yield of dextrose and xylose was seen in sample 3150. After 48 hr about 80% of total sugar within sample 3150 was digested. Sample 3150 was treated with 40% moisture, an ammonia ratio of 2 for 5 min at 90 °C. These process conditions had a dramatic effect on the hydrolysis of rice straw biomass. Initially, Sample 345 had a greater percent yield of dextrose than 3150 but a smaller percent yield of xylose. From 3-48 hr, sample 387 consistently had the smallest yield of both dextrose and xylose for an AFEX treated sample. 387 was treated at 60% moisture for 5 min at 90 °C with an ammonia ratio of 0.5. After 48 hr only about 22% of dextrose and 10% of xylose was obtained for untreated rice straw. Treatment conditions on the hydrolysis rates at 3, 24 hr were inconclusive. With respect to sample 3150 and 387 we concluded that; moderate moisture (40%), higher ammonia ratio (2), smaller reaction time (5min) higher temperature (90 °C) were optimal conditions for the hydrolysis of AFEX treated rice straw. Trends in the Percent Yield Trends in the percent yield of dextrose and xylose versus hydrolysis time are shown in Figure 3.4 and Figure 3.5 respectively. 42 Percent Yield of Dextrose (%) Trends in Percent Yield of Dextrose 100 .L T f I I I I I I I I f I I I r 1 I T I T +8150 -I- Untreated Rice Straw + 811 4 ....7—- - V ...—....-.— 8 + Deiignified Rice Straw at O .,. 4o _ / l i l 20 - F l o b l . . . l l c . . l l . . l I . . - L l . i 0 1O 20 30 40 Hydrolysis Time (hr) Figure 3.4 Trends in Percent Yield of Dextrose 3150 (40% Moisture, Ammonia Ratio 2, 90°C, 5min) B114 (20% Moisture, Ammonia Ratio 1.5, 80°C, 5min) B180 (60% Moisture, Ammonia Ratio 2, 90°C, 5min) 43 50 Percent Yield of Xylose (%) 100 60 80 20 Trends in Percent Yield of Xylose 4 4 a . I . 4 I . 4 r . , —O— 8150 + Untreated Rice Straw 1 +3114 4 _ + Deiignitied Rice Straw .. , J H + l a J 4% l n I n L 1 0 20 30 40 50 Hydrolysis Time (hr) Figure 3.5 Trends in Percent Yield of Xylose B150 (40% Moisture, Ammonia Ratio 2, 90°C, 5 min) B114 (20% Moisture, Ammonia Ratio 1.5, 80°C, 5min) B180 (60% Moisture, Ammonia Ratio 2, 90°C, 5min) The plots (Figures 3.4, 3.5) Show that the rate of hydrolysis is not linear but the rate of hydrolysis is greatest from 0-3 hr. Also, as seen by the plot of 3114 and 3150 the initial rate of hydrolysis (0-3hr) did not necessarily indicate a greater extent of hydrolysis probably because Cd and 1i gnin content of the rice straw affect the hydrolysis rate at different stages of the hydrolysis. Sample 3114 has the same hydrolysis trend as sample 3150 at 3hr but afterwards the rate of hydrolysis decreases drastically for sample 3114. The zero point is not seen because the values were between 03-05% yields of dextrose and between 001% for xylose. Delignified rice straw has a 55% yield of dextrose but only a 25% yield of xylose. The delignification process is moderately effective at improving the yield of dextrose but not the yield of xylose. Only a 25% yield of xylose was achieved. 3.2.3 Statistical Correlation Since rice straw was treated at various conditions, we tried to find a relationship between the initial treatment conditions and the concentration of sugar from hydrolysis studies. The predictor variable (X) was the treatment conditions and the dependent (Y) or response variable was the concentration of either dextrose or xylose. Using regression analysis of 49 samples we tried to find a correlation between the treatment conditions (predictor variables) and the concentration of either dextrose or xylose. The regression analysis used the method of the least squares. The degree of association between the X and the Y was established by 1'2 otherwise known as the coefficient of determination. The 12 may be interpreted as the proportionate reduction of total variation associated with the use of the independent variable X. The larger 1’2 is, the greater the variation of Y is I reduced by introducing the variable X. The closer r2 is to l, the greater is the degree of 45 association between X and Y. An r2 of zero shows no association between X and Y. The 12 from the regression analysis of concentration or percent yield of sugar versus initial treatment conditions are below. 46 Table 3.3 R2 values for Correlation between Treatment Conditions and Sugar r2 Percesnlfglzleld of Demose X ylose Total Sugar 3 hr 0.16 0.15 0.13 24 hr 0.06 0.299 0.05 48 hr 0.027 0.297 0.06 Concesr‘lltgztrron Of Dextrose X ylose Total Sugar 3 hr 0.23 0.37 0.276 24 hr 0.16 0.622 0.33 48 hr 0.21 0.6 0.4 The 2nd, 3rd and 4th columns represent the type of sugar while the rows represent the length of the enzymatic hydrolysis. The intersection of the row and column gives the 1'2 value from the regression analysis done between the initial treatment conditions listed in Table 3.1 and the type of sugar at a specific time point represented at the intersection. The highest correlation was seen between the concentration of xylose at 24-48hr and the treatment conditions. The other correlations were not very significant as concluded by the low r2. There is a reduction of only 60% variability in the concentration of xylose when 49 different treatments were considered. A greater proportionate reduction would have been more useful but Since there is such a large margin of error for the concentration of sugar, it seems that 60% reduction in variability might help to give a rough estimation of the concentration of sugar. 47 Equation for regression analysis: Y = 40 + flrXr + flzXz + .33X3 + .34X4 (3.6) Y: concentration of xylose X1: treatment moisture ((Kg water/Kg dry straw) *100) X2: reaction time (min) X3: reaction temperature (°C) X4: ammonia ratio (lb ammonia/lb dry straw) Bo = y-intercept (g/L) Bl = regression coefficient for X1 ((g/L)/((Kg water/Kg dry straw)*100)) B2 = regression coefficient for X2 ((g/L)/(min)) 03 = regression coefficient for X3 ((g/L)/(°C)) B4: regression coefficient for X4 ((g/L)/(lb ammonia! lb dry straw)) The equation we obtained for predicting the concentration of xylose after 48hr enzymatic digestion was: Yx= -0.234 + 0.000379 X, + 0.0104 X2 + 0.00263 X3 + 0.127X4 (3.7) At 24 hr after enzymatic digestion, the equation is: Yx= -0.2 - 0.0013 X, + 0.01 X2 4» 0.002256 X3 + 0.0993 X4 (3.8) A p-value below 0.05 was only seen for X2 and for X4_ Therefore, for the treatment conditions tested, only reaction time and ammonia ratio have a significant effect on the concentration of xylose while moisture and temperature have a lesser effect. As shown by the positive regression coefficients in the equation for predicting the 48hr xylose concentrations, all the parameters seem to have a positive effect on digestibility of the 48 straw. Only at 24hr after enzymatic digestion does the percent moisture seem to have a negative effect on the concentration of xylose. The apparent discrepancy in the sign of the regression coefficients is explained by the 95% confidence intervals of the regression coefficients. The regression coefficients for X1 and X3 go thru the zero point. Therefore, the regression coefficients can be either positive or negative for X1 and X3. From these results we cannot determine whether treatment moisture and reaction time have a positive or a negative effect on the hydrolysis of the biomass. The plot of the residuals versus each X variable did not show any curves. Therefore, the relationship between the concentration of xylose and the initial conditions is linear. If the residual plots versus each X variable showed a curvilinear relationship, then that particular X should be tested for second or third order relationship. The correlation between xylose and enzymatic hydrolysis at 24-48 hr exists perhaps because alkali treatments are causing the hemicellulose to solubilize. Also, as suggested by Chang and Holtzapple (2000), deacetylation can significantly increase the digestibility of hemicellulose. Hemicellulose is primary composed of xylose sugars and 70% of xylan polysaccharides contain acetyl groups that sterically hinder hydrolysis (Bouveng, 1961). Another factor could be that the explosion of the fibers is causing disruption of the lignin molecules, which allow the enzymes that were previously blocked to adsorb to the substrate. Predicting Ultimate Sugar Yield Since the rice straw takes at least 48 hr to digest fully, we wanted to see if the 48 hr sugar concentrations or ultimate sugar yield could be predicted using the 3 hr sugar 49 concentrations or the initial sugar yield. Figures 3.6-3.8 show the relationships between 3hr and 48 hr sugar concentrations. 50 Predicting Concentration of Dextrose at 48hr 0.8lll'TIIillllrrlrrrrIrr Equation of line: y = 0.333 + 1.01x 0.7} R-square: 0.26 _- 0.6 l I 1 I I I 0.5 l 0.4 Concentration of Dextrose at 48hr (g/L) l l 1 l L l l 0.3 _ #- -l 02llllllllllLLiLiLllllllll 0.05 0.1 0.15 0.2 0.25 0.3 Concentration of Dextrose at 3hr (g/L) Figure 3.6 Predicting Ultimate Yield of Dextrose using Initial Hydrolysis Rate 51 Predicting Concentration of Xylose at 48hr 0.5HTIIIIIIIFIIIIIIIIIITIIIITITTIIII . I A . . S, - Equation of Line. y=0.04 + 2.02x : 0 4; R-square=0.53 . j .c ' _ tn <- .. 4-0 _ (0 m _ tn , - _ 2 0 3 . > >- X .._ - O t g 0.2— — .... - g . c .. 8 . C .1 — .. o O _ , O _ . . 0 P 1 l I i l l l i 1 1 1 l l l l l l l l 1 1_LJ I l l l l l I 1 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 Concentration of Xylose at 3hr (g/L) Figure 3.7 Predicting Ultimate Yield of Xylose using Initial Hydrolysis Rate 52 Predicting Concentration of Total sugar at 48hr 1.2llll[IllIllllTIlllllllTTllllllTj’TTrllll Equation of Line: y=0.38 + 1.3x R-square = 0.34 Concentration of total sugar at 48hr (g/L) 02 1111l1144111LLILJ_LllL;11LLL11‘LL1L1'11:! I 0.1 0.15 0.2 0.25 0.3' 0.35 0.4 0.45 0.5 Concentration of total sugar at 3hr (glL) Figure 3.8 Predicting Ultimate Yield of Total Sugar using Initial Hydrolysis Rate 53 An r2 of 0.5273 was found for predicting the 48 hr concentration of xylose using the 3 hr xylose concentration ( Figure 3.7). Using the concentration of xylose at 3 hr reduces only 52.73% of the total variation of the concentration of xylose at 48 hr. The concentration of xylose at 3 hr was the best predictor but an r2 of 0.5273 has a large amount of scatter. Equation for regression analysis: Y: ,80 + ,BIX (3.9) Y: concentration of xylose at 48 hr X: concentration of xylose at 3hr Bo = y-intercept (g/L) [31 = regression coefficient for X (unitless) Equation of line: Y = 0.04133 + 2.02 (X) (3.10) Predicting the concentration of dextrose at 48 hr and concentration of total sugar at 48hr using 3hr sugar concentrations, gave an r2 of 0.26 and 0.34, respectively. They show a lesser dependence of Y on X and are probably not acceptable. One of the reasons that there is little correlation between ultimate and initial sugar yields maybe because some of the samples have not been fully hydrolyzed. Also, characteristics of the rice straw such as crystallinity of cellulose and the content of lignin could differently affect the rate of hydrolysis at different stages of the hydrolysis. 54 3.3 Summary AFEX treatment of rice straw has many advantages as shown by the enzymatic hydrolysis studies. AFEX treatment improves the digestibility of the rice straw. The enzymatic studies show that the AFEX treatments can maximize digestibility of rice straw up to 80% as the sugar yields for sample 3150 indicate. Also, not a significant amount of residual ammonia was found on the rice straw thus, the pH of the rice straw was not affected. Consequently, no washing of the biomass was required. Since washing is not required there is no loss of biomass. The main disadvantage of the process seems to be that there is not a uniform treatment of the rice straw along the reactor. Even though the process has to be further modified, still most of the AFEX treatments significantly increase the digestibility of the rice straw. The higher 1’2 of the correlation between the xylose concentration and the treatment conditions seems to support the fact that the treatment conditions especially reaction time and ammonia ratio are affecting the hydrolysis of hemicellulose either by solubilization, deacetylation or by disruption of lignin. The treatments are also, significantly increasing the rate of digestion of cellulose but we need to further study the effect of Cd and lignin content on the hydrolysis rates before we can make any conclusions. Initial hydrolysis rates of dextrose could not predict the ultimate yields of dextrose but initial hydrolysis rate of xylose maybe be able to predict ultimate yield of xylose. The delignified rice straw had a moderate increase in yield of dextrose but not a significant yield in xylose. Therefore, 1i gnin content definitely affects the digestion of cellulose. To fully understand what parameters are affecting hydrolysis of the AFEX treated biomass, we are going to 55 next determine the Cd and lignin content of the samples. We will further analyze the Cd and lignin content of the samples to determine whether any of these factors correlate with digestibility. 56 4 CRYSYTALLINITY INDEX 4.1 Introduction 9 When a beam of monochromatic x-ray radiation is directed at a crystalline material, one observes reflection or diffraction of the x-rays at various angles with respect to the primary beam. The relationship between the wavelength of the x-ray radiation, the angle of diffraction, and the distance between each set of atomic planes of the crystal lattice is given by Bragg’s equation. 71. = 2dsin6 (4.1) From Bragg’s equation one can calculate the interplanar distances ((1) of the crystalline material. The interplanar spacings depend on the geometry of the crystal’s unit cell while the intensities depend on the type of atoms in the crystal and the location of the atoms in the fundamental repetitive unit cell. Rice straw contains cellulose, which is partially crystalline and partially amorphous. Occurrences of amorphous or noncrystalline regions within the cellulose are a consequence of strain distortions between the microfibrils. The crystallinity index is a measure of the relative amounts of the crystalline and amorphous regions in cellulose (Segal et al., 1959). 57 4.2 Experimental 4.2.1 Methods The same powdered rice straw used for the hydrolysis studies was also used for the cystallinity studies. Crystallinity of the rice straw is measured using X-ray diffraction. Crystallinity index (CrI) was measured using the Rigaku Rotaflex 200B. The sample was placed vertically and analyzed using the horizontal goniometer. Triplicate samples were scanned at 1°lmin from 2 0 =10°- 26° with a step size of 005°. The crystallinity index was determined as defined by Segal et al. (1959): Iam)*100 lam Crystallinity Index (Crl) = ([002 - (4.2) 1002 is the intensity of the peak at 22.8° Iam peak is the intensity of the peak at 18° 58 4.3 Results and Discussion 4.3.1 Trends in CH The intensity of the replicate x-ray diffraction plots varies quite significantly but the crystalline and amorphous peaks remain proportionate. The CrI of the sample is reproducible with a replication error of i 0.3. The crystallinity indices of the various AFEX treated samples are shown in Table 4.1. 59 Table 4.1 Crystallinity Index of the AFEX Treated Rice Straw ID I... Ion: CHI B103 81 113 28.31858 B1 04 200 319 37.30408 B105 168 255 34.11765 B1 08 583 995 41 .40704 B11 94 129 27.13178 B111 208 311 33.11897 B114 125 180 30.55556 B120 124 173 28.3237 B123 221 313 29.39297 B126 209 317 34.0694 B135 1 13 163 30.67485 B1 38 224 333 32.73273 B1 4 187 257 27.23735 B1 44 1 15 163 29.44785 B1 47 99 1 58 37.341 77 B150 146 205 28.78049 B1 53 228 353 35.41 076 B1 56 142 209 32.05742 B1 59 175 279 37.27599 B1 62 160 250 36 B165 160 236 32.20339 B1 68 268 439 38.95216 B17 76 111 31.53153 B1 77 242 387 37.4677 B1 80 132 21 8 39.44954 B186 175 261 32.95019 820 230 335 31 .34328 B24 156 240 35 828 177 273 35.1 6484 832 236 328 28.04878 '60 Table 4.1 (Cont’d) ID I... Im cm 835 109 155 29.67742 838 87 115 24.34783 841 174 262 33.58779 845 109 151 27.81457 851 142 212 33.01887 854 249 364 31.59341 857 281 374 24.86631 860 194 249 22.08835 863 139 21 37.10407 866 91 112 18.75 869 80 121 33.8843 872 173 258 32.94574 B75 97 134 27.61194 878 234 317 26.18297 88 85 129 34.10853 882 349 599 41.73623 884 189 246 23.17073 890 82 108 24.07407 893 676 1084 37.63838 C172 310 543 42.90976 0184 125 219 42.9237 049 81 116 30.17241 untreated flee straw 226 381 40.68241 Deli nified angst,“ 180 340 47.05882 61 The CrI of untreated rice straw was 40.7, which was slightly higher than a CrI of 35 obtained for untreated southern pine pulp by Bertran and Dale (1985). This result is expected because the composition of rice straw and pine pulp might vary. The results shown in Table 4.1 indicate a definite impact of AFEX treatment on CrI of rice straw. The treatments caused a variation of Cd from a minimum of 18.75 to a maximum of 42.9. Most of the AFEX treatment conditions caused a decrease in the CrI of the treated samples. Sample 8150 with the highest yield of dextrose and xylose, had a significantly lower CrI (28.8) than untreated rice straw but it did not have the lowest CrI as expected. On the contrary, sample B66 (Figure 4.3) with the lowest CrI (18.75) and sample B60 with 8 Cd (22) had only a 40% percent yield of dextrose (42%) and xylose (30%). Although samples B150, B66, B60 were all treated at 40% moisture, sample 3150 was treated at a higher temperature (90 °C) and ammonia ratio (1.5) for lesser time (5min). The low CrI of B66 and B60 along with low yields of sugar indicate that the pretreatments are probably increasing the amount of amorphous cellulose by decreasing the crystallinity of the cellulose structure but lower rates of saccharification are probably due to reducing access to the substrate (cellulose and hemicellulose) of the celluclast enzymes by lignin. For samples with high lignin and low CrI the samples, high rates of saccharification should take place within 3 days because the few adsorbed enzymes should effectively digest the substrate due to low crystallinity (Chang and Holtzapple, 2000). If the enzymatic hydrolysis duration were longer than 48 hr for our experiments, we probably would have seen higher rates of hydrolysis for samples B60 and B66. Other samples had similar CrI as 3150 but they had much lower yields of sugar. 62 Samples C172, C184, B82, B108,.had a slightly greater CrI (41.7-42.9) than the untreated rice straw. The X-ray diffraction profile for sample B108 ( Figure 4.1), which was treated at 60% moisture, 1.0 ammonia ratio for 10 min at 90 °C, reveals some sharp peaks. These sharp peaks indicate that the crystalline cellulose structure is highly ordered and therefore very resistant to enzymatic hydrolysis. As expected, hyrolysis of sample B108 resulted in very low yields of dextrose (29%) and xylose (26%). The profile for sample C172 (Figure 4.2) and untreated rice straw (Figure 4.4) had a similar appearance to 3108 but did not contain the sharp peaks. Samples C172 and C184 had a 30—50% yield of dextrose but a higher percent yield of xylose (60-74%). Even though the Cd is slightly higher than for untreated rice straw, all the samples had higher yields of sugar than untreated rice straw. The samples with the lowest CrI were treated at 40% moisture, 1-0.75 ammonia ratio for 10 min at 80 °C while the samples with the highest CrI were treated at 60%, 1.5 ammonia ratio at 80 or 90°C for 10min. The difference of treatment between the highest and lowest Cd is the moisture and ammonia ratio. The samples with the highest CrIIhave similar yields of dextrose as samples with the lowest CrI but the yield of xylose is much higher for these samples. Perhaps the treatment conditions for the samples with the highest CrI are causing removal of amorphous hemicellulose and therefore the xylose yields for these samples are increasing. Hemicellulose is primarily composed of xylose sugars. Although cellulose is the largest contributor to the measured amount of crystallinity, removal of amorphous lignin, amorphous cellulose and hemicellulose from the cellulose matrix can cause the CrI to increase (Gharpury et al., 1983). We found that complete delignification of rice straw can also increase the CrI. Delignified rice straw 63 had a similar profile as untreated rice straw but with a higher proportion of the crystalline to amorphous peak (CrI of 47). Deli gnification process using the permangante method (Van Soest and Wine, 1968) removed the lignin but it did not affect the carbohydrate composition of the rice straw (Edwards and Mackney, 1939) as a result, the structural cellulose and hemicellulose were not affected. The problem with the delignification process is that the capillary structure of cellulose collapses and the physical structure changes when the cellulose is air-dried from the water-swollen state (Caulfield and Steffes, 1969; Howsman and Marchessault, 1959). Therefore removal of just lignin can cause the CrI to increase but also the pretreatment can affect the crystalline structure of cellulose microfibers and change the CrI. For samples C172 and C184, probably there is some removal of amorphous lignin and cellulose, indicated by the higher yields of glucose than untreated rice straw, but since the samples are increasing in xylose digestibility we concluded that there is also removal of amorphous hemicellulose from the crystalline structure and increased solubilization of hemicellulose. Intenstiy (arbritary units) Profile of Sample B108 1000_‘"["T]"'['II]III]rT[1111111 900 j- } -, 800 3 -j 700 :— -§ 600 _ I 11111! 500; j ‘ 400 IIIIUU 1‘11 300 _ ZOO-1441114114411rnLgrnlrrrlrrLJrrr 1012141618 20 22 24 26 Degrees Figure 4.1 X-ray Diffraction profile of Sample B108 (60% moisture, 10min, 90°C, ammonia ratio of 1.0) 65 Intensity (arbritary units) 600 _.,. 550 h 500 450 400 ' 350 Z 300 l 250 I 200 Profile of Sample C172 1 T I I T W I I I l l I I I T j I I 1T l’fi T I j r ‘1 : 1 1 b_ l l l 1 1 l 141 l 4 l I l l l l 4 1 L l l J l J 1 1_l 1 l J_L 10 12 14 16 18 20 22 24 26 Degrees Figure 4.2 X-Ray Diffraction Profile of sample C172 (60% moisture, 10min, 80°C, ammonia ratio of 1.5) 66 Intensity (arbritary units) 130_ 110’ 100: 70 50' Profile of Sample 866 120: IIlllIIIIrIIIIIIIIIIIIfiIWIITIT 90; 80: 60- I I I I ' l ngllAlllngjlll11111441111111114 10 12 14 16 18 20 22 24 26 Degrees Figure 4.3 X-ray Diffraction profile for sample B66 (40% moisture, 10min, 80°C, ammonia ratio of 1) 67 Intensity (arbritrary units) Profile of Untreated Rice straw. 4OOfiTIIIIIIIIIIIII‘TIIFTTTjIj—IfiTrj 350% 150 111 11 1 1 1111 [1 lllLll 111111 1111 11‘ 10 12 14 16 18 20 22 24 26 Degrees Figure 4.4 X-ray Diffraction profile of untreated rice straw 68 4.3.2 Statistical analysis Since crystallinity of cellulose has an impact on digestibility we tried to model CrI of the rice straw versus concentration of sugar. Linear regression was done on KaleidaGraph using the method of least squares. The predictor variable (X) was the crystallinity index (Cd) and the dependent (Y) or response variable was the concentration of either dextrose or xylose. Using regression analysis of 49 samples we tried to find a correlation between the CrI (predictor variables) and the concentration of either dextrose or xylose. The degree of association between the X and the Y was again established by r2. An 1’2 closest to 1 was preferred. The closer the r2 is to 1, the greater the reduction of variability of the model. Figures 4.5-4.10 show the relationship between Cd and concentration of sugar. 69 Concentratlon of Dextrose (glL) Crl vs Concentration of Dextrose at 3hr 0'3II"I""r""]FTtrlfirvrjrr1r " e 0 Equation of line: y = 0.1793-0.000914x R-s uare=0.0081 0.25 — q o e e e ” e 0.2 T 0 O . . . e o O I 1- . . 0.15 _ . . - e e g " C C . b e ’ e e 0.1 '— . O. .0 .. O . . . e e . e 0.05 l 4 l l J J 1_J l I 1 l L J l l l l 1 LL 1 L i l l 1 L 15 20 25 30 35 40 Crystallinity Index (%) Figure 4.5 Cd versus concentration of dextrose at 3hr. 70 Concentratlon of Dextrose (glL) Crl vs Concentration of Dextrose at 24hr 0'7 I I I I I r I I I I I I I I I I I I I T I I f I I I l W 1 Equation of line: y = 0.513-0.00459x ' R- uare=0.068 0.6- 5“ O. " O 0.5- . . b D 0.4— 0.3" " O D O O '- 0 O O D 0.2’ . I . 0.111111411111111L11111111111411 15 20 25 30 35 40 Crystallinity Index (%) Figure 4.6 Cd versus concentration of Dextrose at 24hr 71 Concentration of Dextrose (glL) Crl vs concentration of Dextrose at 48hr 0.8 b I I I .I T I j f rj I I j I I I T I I I I I I I I I I I I d : Equation of line; y = 0.48 + 7.33e-5x ' - 0-7 f R-square = 1.31e-5 —‘ ’ e I . e 0.6 .— . . . O _. - . . O . -1 ' 9 0 . e e ’ e 005 — . . -4 O _ : 0 O " es s . . . . e . e e ’ e 0.4 :— . . O 0 ~— 8 0 . 0 . e .- 0 .. 0.3 f j - . - 0.2 - _ I o - : 1 l I L 1 1 0.1 A l l 1 L 1 1 1 1 1 1 1 1 1 1 4 1 1 1 1 1 1 1 15 20 25 30 35 40 45 Crystallinity Index (%) Figure 4.7 Cd versus concentration of dextrose at 48hr 72 Concentratlon of Xylose (glL) Crl vs Concentration of Xylose at 3hr 0'18 I I I I I I I T I I I I I I rT—r‘ I—rl I I I I I I I I— T Equation of line: y = 0.0554 + 0.00105x o_1 6 : R-squarc = 0.02285 0.14— - ,o . r C O 012- e - 0.1— p 008_ - h 0.06- ° ' ' ' P ' o 0 r .0 C. C O 0.04” i. r L. 0 . 002111111IJIIIiIPIIIIILLIIIIIILL . 15 20 25 30 35 40 Crystallinity Index (%) Figure 4.8 Cd versus concentration of xylose at 3hr 73 Crl vs Concentration of Xylose at 24hr o-4_""I"‘—El""l""l'7"l"" : Equation of line: y = 0.078 + 0.0028x 0.35 - R-square = 0.0379 _ IT‘F L .1 h 4 — —4 0.3 I 111111 p h 0.25 b P P N r 0.15 TII a n ‘ Concentratlon of Xylose (glL) IIIIII l 0.05 b p b ‘ o1111[11111m111111L4111rilLLLr 1 5 20 25 30 35 40 45 Crystallinity Index (%) Figure 4.9 Crl versus concentration of Xylose at 24hr 74 Concentration of Xylose (glL) Crl vs Concentration of Xylose at 48hr 0.5 I I I I I fi‘I I I I I I I I I T 1 I I I I I I I I I T I T E Equation of line: y = 0.041 + 0.0057x . R-square = 0.092 0.4 — - ' ' — .- C C C C C " _ C C - C 0.3 '- — T 0.2 — ... r .1 0.1 '_ . _ C o F 1 1 1 1 1 L 1 1 1 L 1 1 1 1 J 1 1 1 1 I 1 1 1 1 1 15 20 25 30 35 40 Crystallinity Index (%) Figure 4.10 Crl versus concentration of xylose at 48hr 75 45 The r2 values (Table 4.2) obtained from the regression analysis were very small and almost close to zero. Therefore we can conclude that there is no correlation between Cd and the concentration or percent yield of xylose or dextrose of each sample. This is consistent among all the plots. Table 4.2 R2 Values for Correlation between Cr] and Sugar '2 Percesrzglzreld 0’ Dextrose Xylose Total Sugar 3 hr 0.026 0.0057 0.0063 24 hr 0.119 0.014 0.036 48 hr 0.03 0.06 0.0002 Concentration of Dextrose Xylose Total Sugar Sugar 3 hr 0.0081 0.0228 8.42E-05 24 hr 0.068 0.0379 0.0037 48 hr 1.31 E05 0.092 0.03 An interesting point to note is that the concentration of dextrose versus CrI gives a negative slope while the concentration of xylose versus CrI gives a positive slope. Only at 48hr dextrose concentration (Figure 4.7) does the slope become horizontal. The overall trend shows an increase in the concentration of xylose as the Crl increase. On the contrary, a decrease in the concentration of dextrose can be seen as the CrI increases. We can probably conclude that some treatments increase the CrI of biomass by removing amorphous 1i gnin and hemicellulose from the crystalline cellulose (Gharpuray et al., 1983) thus, increasing the susceptibility of hemicellulose to digestion. Therefore the concentration of xylose increases with increasing CrI. On the other hand, the concentration of dextrose decreases with increasing Crl. Cd is a relative proportion of amorphous to crystalline cellulose. The higher the Cd is, the greater the amount of crystalline cellulose. The rate of enzymatic hydrolysis depends on the crystallinity of the 76 substrate. The accessible amorphous portion of the cellulose degrades much faster than the less accessible crystalline regions (Fan et al., 1980; Baker 1973). Therefore, as Cd is increasing the yield of dextrose is decreasing. We are not able to predict the relationship between Cd and digested sugars because not only Crl but also probably the lignin content of the rice straw is being affected by the treatments. Also, the error from the hydrolysis studies reduces our ability to find a correlation. Therefore, Cd is not the only factor being affected by the treatments and it not the only factor affecting the saccrification rates of sugars. Although the AFEX treatment conditions are affecting the CrI, the association between treatment conditions and Cd is not very significant. Regression analysis using the least square methods gave us an r2 of 0.286. There is no correlation because AFEX treatment not only affects the crystallinity cellulose, but it probably removes lignin and also increases surface area thus increasing the susceptibility of enzymes hydrolyzing the cellulose and hemicellulose (Holtzapple et al., 1991). These aspects of the rice straw especially the li gnin content and Crl vary with each treatment therefore not all treatments may affect the crystallinity of cellulose while other treatments may dramatically change it. 4.4 Summary Most of the AFEX treatments decrease the CrI 0f the rice straw samples but a few of the samples had a higher CrI than the untreated rice straw. N 0 correlation was found between the CrI of the treated rice straw samples and the concentration or yield of dextrose and xylose from 3, 24 and 48hr after enzymatic digestion. The treatment conditions affected the CrI of the rice straw but no trend could be found. These results 77 lead us to make some conclusions. First of all, the Cd is not an accurate measure of the cellulose crystallinity because of the presence of amorphous lignin and hemicellulose. Also, CrI does not completely determine enzymatic digestibility and therefore, Crl is not a good predictor of digestibility. 78 5 FLUORESCENCE 5.1 Introduction Fluorescence emission occurs when a molecule emits light after absorbing energy from a radiation source. Absorption of energy from the radiation source can promote an electron to an unstable higher energy level. The excited molecule then undergoes internal conversion or collisions with the surrounding molecules, which causes the molecule to lose energy. The loss in energy in turn causes the molecule to descend to the lowest vibrational energy level of the excited state. From this excited state the electron can return to its original energy level by emitting radiation (fluorescence). Usually molecules with double or conjugated double bonds fluoresce after excitation. The energy transfer for these molecules usually occurs between a donor (high energy bond 1m“ type singlet excited state) and an acceptor (carboxyl or carbonitrile groups). The energy transfer causes the molecule to lose energy. Lignin, a complex polymer within the plant cell wall, is primarily composed of phenol rings. These phenol rings are the main source of autofluorescence for lignin. Lignin fluoresces with a broad maximum at a wavelength of 455nm when excited at a wavelength of 350nm. Molecules with similar structures to lignin can also fluoresce within the lignin structure or in the solid sample matrix. Lignocellulosic biomass contains many fluorophores, which fluoresce in the same region (Castellan er al., 1994). Cellulose, a carbohydrate polymer in the plant cell, is also auto fluorescent at a maximum of 455nm when excited at 350nm (Olmstead and Gray, 1997) but Castellan er al. (1994) found that wood fluorescence was mainly due to lignin and not from cellulose. Although the interpretation of fluorescence spectra of lignocellulosic material is complex (Beyer et 79 al., 1995; Olmstead and Gray, 1993; Tylli et al., 1995, Billa et al., 1999) found a good correlation between fluorescence data and the kappa number along With the extracted lignin content of pulps made from various raw materials. Therefore, fluorescence is an inexpensive, rapid, specific, non-destructive and non-intrusive, method that can be used to characterize lignin content within rice straw. Fluorescence detection of solids is complicated. Unlike liquids, the fluorescence of solids cannot be detected at a 90-degree angle. Measuring solids at a 90—degree angle may not give reproducible or detectable data due to self-absorption and complete attenuation of the beam. Therefore, a method called front face detection is used for analysis of solid and other opaque materials. Front face detection is a technique where light is focused on the front face of the sample and then fluorescence emission is collected from the same region at an angle that minimizes reflected and scattered light. 5.2 Experimental 5.2.1 Methods Fluorescence spectra wererecorded using a SPEX-3 Fluorolog. Auto emission spectra were obtained at an excitation wavelength of 350 nm with an interval of 0.5nm. The excitation and emission slit widths were set at 3 and 5 nm, respectively. The solid sample holder was filled with the powdered sample and was held in place with a quartz cover slip. The mode of detection was set at front face. Each sample was measured three times. All the fluorescence data were normalized. The minimum point on the fluorescence graph was usually at the baseline. (Data point — minimum data point) (maximum data point — minimum data point) Normalized data = (5.1) 80 The area under the fluorescence peak from 400-650nm was determined by using the trapezoidal rule as shown in Equation 5.2. The area of each trapezoid was added to determine the total area from 400-650nm. 0.5 * ( Wavelengh 2 - Wavelengthr) ( N ormalizedIntensitye - N ormalizedIntensity 1) Area: (5.2) 81 5.2.2 Statistical analysis methods Multivariate Linear Regression (MLR) Multivariate Linear regression (MLR) analysis was done using the statistical package on MS Excel. MLR regression was done between the chosen fluorescence data and the concentration and percent yield of sugar. The predictor variable (X) was the fluorescence data and the dependent (Y) or response variable was the concentration or percent yield of either dextrose or xylose. Only the intensity at a specific emission wavelength or a few specific emission wavelengths was used for the MLR correlation studies. Also, the area of the corrected fluorescence data was used to find a correlation. Using 51 samples we tried to find a correlation between the fluorescence data (predictor variables) and the concentration or percent yield of either dextrose or xylose. The degree of association between the X and the Y was again established by r2. An 1'2 closest to l was preferred. The response variable (Y) did not depend on predictor variable (X), if the r'2 was close to zero. 82 Principal Component Regression Modeling The spectra were normalized and then mean scaled (covariance about the mean) by subtracting each spectrum from the mean spectrum for PCR regression modeling. The fluorescence data was also analyzed using Principal Component Regression (PCR) on MatLab using the code shown in Appendix Table 8.7. All 501 fluorescence data points were separated into principal components using the orthogonal factors in a stepwise regression. The MLR analysis was done between the chosen principal components and the concentration or percent yield of sugar. Also, MLR analysis was done between the sugar values and a combination of fluorescence principal components and CrI. The regression coefficients were set at 99% confidence interval. This confidence interval allows us to predict the limits of the each regression coefficient. The degree of association between the X and the Y was again established by r2. 83 5.3 Results and Discussion When excited with alight at a wavelength of 350nm, the AFEX treated rice straw samples gave an emission profile ranging from 400-650nm, with a peak around 480nm. The fluorescence was detected from 400-650nm because the emission profile always occurs at a lower energy than the excitation wavelength due to energy loss. Also, the first Rayleigh peak occurs at the excitation wavelength (350nm) and the second Rayleigh peak occurs at twice the excitation wavelength (700nm). These Rayleigh peaks are anomalies and occur due to interference factors. Therefore, the emission profile was set between 400 and 650nm. An excitation of 350 nm was chosen because the highest fluorescence intensity of treated wheat straw was measured at 350nm (Billa et al., 1999) and also fluorescence measurement of mechanical paper pulps by Tylli et al. (1996) was done at 350 nm. Although wheat straw has been shown to fluoresce when excited at other wavelengths, the maximum intensity was seen at 350 nm (Billa er al., 1999). Fluorescence was measured with an entrance slit width of 3 nm and an exit slit width of 5 nm as using a larger slit width increased the fluorescence intensity beyond the linear range. Decreasing the slit width further minimized the intensity beyond what was necessary. We measured the fluorescence of powdered delignified rice straw (Figure 5.1) to see if the sample could be used as a blank but the fluorescence of delignified rice straw had the same intensity as powdered untreated rice straw with a few minor differences. Although the delignified sample did not have reduction of overall fluorescence, there was reduction of fluorescence in the region of 480-500nm and at 527nm. One would have expected to see complete elimination of fluorescence but this did not happen. Plant cell 84 walls contain other products such as cutin, suberin, sporopollenin and cellulose, which fluoresce when excited with light (Hartley, 1853). Also, molecular structures that are not generally known to fluoresce may do so in the lignin structure or in the solid pulp matrix (Lamola et al.,l965). It seems as though fluorescence of the samples may not be entirely attributed to lignin but other components of the rice straw may contribute to the fluorescence profile. Another reason for the intensity of the fluorescence emission spectra may be the effect of the treatment (T ylli et al., 1996) using the permanganate method. Upon treatment the rice straw became very fluffy and even sieving did not help to retain its powdery texture. Also, the color changed from dark brown to a white shade. After treatment the particles easily adhered to one another. All these factors could have changed the refractive index and therefore affected the fluorescence intensity. A blank was not utilized because an accurate blank should have the same properties of rice straw except without lignin and delignified rice straw did not serve the purpose. 85 Intensity (arbitrary units) 1.2 0.8 0.6 0.4 0.2 -0.2 —°— Delignifled Rice Straw -°— Untreated Rice Straw Effect of Dellgniflcation on Rice Straw f‘r I I T l I I I 1 I IfI I r I I I I I I I I I I fiI T fT T I r 1 I 1 I 1 1 J 350 400 450 500 550 600 650 Figure 5.1 Effect of Delignification on fluorescence Spectra Wavelength (nm) 86 700 5.3.1 Reproducibility of fluorescence Profiles The effect of particle size on reproducibility was studied. Some of the rice straw was retained on 40-mesh (0.35mm openings) screen while some rice straw was sieved through an Opening of 0.106 mm (150 mesh). Sieving through the ISO-mesh screen gave us the powdered sample. Figure 5.3 gives the fluorescence profiles of the two different particle size. Figures 5.3 and 5.4 show the reproducibility of the two different particle sizes. 87 Intensity (arbitrary unlts) —e— 3165 -9— sample replaced -0 - new sample 1'2 r r I 1 1 I I I I ' ' i j I r I T T I I I I r I r r a j I Y r 1 " _ F q r- -1 0.8 b —1 0.6 ” _ 0.4 F _ P d h -1 0.2 i‘ _ p .4 F . o — —1 f . -02 1 1 LJ_J 1 1 1 1 I 1 4 1 1 1 1 1 m_1 J 1 1 L 1 I 1 1 1 1 j 1 1 1 1 350 400 450 500 550 600 650 Wavelength (nm) Figure 5.2 Fluorescence Profile for Powdered (106nm) Rice Straw 88 1.2IIIIIITIIIIIIIIIIIIIIIITTIIIIIIIII 1 ... _ g 0.8 :_ r IM‘ L g h u MK 1 3' ll It 5 \l g 0 6 — J1 “"11 - 3 ,1 *1, >1 8 r ““1 § 0 4 - k.” - E 1. oAlllllllnglliJLilJmlllilJll1111 1111 350 400 450 500 550 600 650 700 Wavelength (nm) Figure 5.3 Fluorescence Spectra Reproducibility of 035mm Size Rice Straw 89 —-—Powdered sample -9- Larger particle 12 V 1 Y V—Ii ‘ I I T I I 7 V I l I I Y Y I 7 f' I I 1 I I I' [Tfir 17f 0.8" 06* C Intensity (arbitrary units) 0.4 ” 0.2‘ 0 ...1 350 400 450 500 550 600 650 700 Wavelength (nm) Figure 5.4 Effect of Particle Size on the Fluorescence Spectra 9O As illustrated by Figure 5.2, the amount of lignin did not vary with duplicate spectra of powdered rice straw (106um). However, only the region from 400-470 nm lacked reproducibility. This is probably due to scattering of stray light. The larger particle rice straw (0.35mm) did not provide satisfactory reproducibility as seen in Figure 5.3. This discrepancy is possibly due to the changing refractive index caused by the inconsistent packing of the larger particle rice straw. Figure 5.4 illustrates the effect of particle size on fluorescence spectra. The larger particle size had a different emission profile than the powdered rice straw for sample 8165. The two plots show similar peaks but the intensities are different. One reason for the difference could be again due to the refractive index, which varies with particle size. Another explanation for the variation in fluorescence profiles could be due to grinding which exposes more of the lignin to the 350nm wavelength light. Since the error is significant in Figure 5.3, the variation is probably caused by the refractive index rather than the difference in lignin. In view of the fact that the powdered rice straw gave a more reproducible fluorescence profile, all the data were collected using rice straw that was sieved through 106 um openings (ISO-mesh) to give a homogeneous surface. 5.3.2 Trends in Fluorescence Spectra We wanted to measure the variation in the fluorescence spectra of samples with varying concentrations of sugars (dextrose and xylose). These concentrations of sugars were obtained from the earlier hydrolysis studies performed on the AFEX treated rice straw samples. The greatest concentration of sugar was found in sample 8150 while the lower yields of sugar were found in untreated rice straw and sample B108. The variability of the fluorescence profiles among the samples is shown in Figure 5.5. 91 —'°—B150 —°— C184 'B— 8108 1.2 T_I T I I I I I I I I I I I TfI I TI I I I I I I I I I I Ifi T I 0.8 r Intensity (arbltrary unlts) c o» I 350 400 450 500 550 600 650 700 Wavelength (nm) Figure 5.5 Effects of Different AFEX Treatments on Fluorescence Profile of Rice Straw B150 (40% moisture, 5min, 90°C, ammonia ratio of 2) B108 (60% moisture, 10min, 90°C, ammonia ratio of 1) C184 (60% moisture, 10min, 90°C, ammonia ratio of 1.5) 92 The dissimilarity among the samples is seen at 400nm, 483nm and at 527nm but the difference at 400nm was also seen in the reproducibility plot (Figure 5.2). The intensities at 400, 483, 527nm and the Overall area from 400-650 nm for, each sample are shown in Table 5.1. 93 Table 5.1 Fluorescence data peak peak peak area ID 400nm 483nm 527nm 400-650 B103 0.540 0.933 0.450 109.1 B105 0.689 0.967 0.415 1 10.6 B108 0.651 0.875 0.433 102.8 B11 0.541 0.963 0.448 1 10.7 B1 11 0.701 0.964 0.422 111.5 B114 0.559 0.974 0.493 1 13.5 B120 0.534 0.983 0.519 115.2 B123 0.598 0.959 0.502 1 15.3 B126 0.487 0.985 0.547 116.6 B135 0.560 0.979 0.540 1 16.5 B138 0.497 0.993 0.489 1 1 1 .1 B14 0.538 0.965 0.501 1 16.1 8141 0.570 0.966 0.487 1 13.8 B144 0.545 0.994 0.470 1 10.4 B147 0.613 0.981 0.460 113.9 B150 0.618 0.982 0.481 1 15.7 B153 0.705 0.941 0.444 116.8 B156 0.530 0.925 0.474 1 17.3 B159 0.700 0.956 0.473 1 15.5 B162 0.616 0.958 0.498 1 16.4 B165 0.669 0.975 0.425 1 1 1.2 B168 0.605 0.954 0.482 113.5 B17 0.537 0.935 0.478 1 12.8 B177 0.737 0.954 0.451 1 14.7 94 TABLE 5.1 (Cont’d) peak peak peak area ID 400nm 483nm 527nm 400-650 8180 0.556 0.954 0.365 99.4 B186 0.833 0.973 0.408 105.7 82 0.343 0.992 0.463 103.6 820 0.654 0.981 0.438 114.7 824 0.867 0.964 0.388 1 16.5 828 0.917 0.914 0.386 113.2 832 0.490 0.966 0.456 108.5 835 0.558 0.947 0.464 1 1 1.5 845 0.452 0.942 0.478 109.2 85 0.422 0.960 0.500 1 10.0 851 0.533 0.949 0.470 1 13.6 854 0.513 0.924 0.483 1 10.0 857 0.749 0.952 0.466 1 17.0 860 0.810 0.946 0.435 1 14.2 863 0.753 0.957 0.387 94.5 869 0.498 0.960 0.471 1 12.3 872 0.757 0.959 0.466 1 15.7 875 1.000 0.848 0.359 105.7 878 0.564 0.973 0.478 1 12.5 88 0.460 0.978 0.471 1 10.6 884 0.654 0.950 0.501 1 16.7 890 0.490 0.959 0.436 104.5 C172 0.817 0.930 0.444 116.0 C184 0.933 0.920 0.435 121.7 C49 0.569 0.938 0.469 1 10.8 rice straw 0.756 0.900 0.374 1 10.5 Delignified Rice Straw 0.75 0.83 0.356 1 15.4 95 5.3.3 Statistical Analysis Multiple Linear Regression between fluorescence data and sugar values Statistical analysis using multivariate linear regression (MLR) was done to find a correlation between the fluorescence data (Table 5.1) and the concentration of sugars from 3-48hr digestion studies. The r2 values for the regression studies are shown below. Table 5.2 R2 values from MLR Analysis of fluorescence data and Conc. of sugar Intensity at Concentration Intensity Intensity Intensity at 400nm, 483nm Area (400- of dextrose at 400nm at 483nm 527nm and 527nm 650nm) 3hr 0.02 0.08 0.135 0.174 0.04 24hr 0.05 0.171 0.293 0.37 0.068 48hr 0.008 0.07 0.07 0.135 0.17 Intensity at Concentration Intensity Intensity Intensity at 400nm, 483nm Area (400- of xylose at 400nm at 483nm 527nm and 527nm 650nm) 3hr 0.124 0.0002 0.004 0.294 0.116 24hr 0.07 0.01 0.035 0.327 0.14 48hr 0.05 0.003 0.006 0.21 0.16 The intersection of the column (intensity or area from the fluorescence data) and row (type of sugar at the specific time point) are the 1'2 values. The r2 values were close to zero for most of the regression studies. Therefore, the correlation between the sugar concentration and chosen fluorescence data is very poor using the MLR method. The highest 1'2 as seen above is given for the correlation between the 24 hr dextrose concentration and the intensity at 400, 483 and 527nm but the r2 is still not significant. The MLR analysis studies between percent yield of sugar and fluorescence data are shown below. Again the 1‘2 values are close to zero and the correlation between 96 percent yield of sugar and fluorescence data is not significant. The highest 1‘2 was for percent yield of xylose at 48hr and the intensity at 400, 483 and 527nm. Table 5.3 R2 values from MLR Analysis of fluorescence data and Percent Yield of Sugar Intensity at Percent Yield Intensity Intensity Intensity at 400nm, 483nm Area (400- of dextrose at 400nm at 483nm 527nm and 527nm 650w 3hr 0.054 0.036 0.077 0.085 0.018 24hr 0.023 0.036 0.072 0.079 0.066 48hr 0.0008 0.0093 0.01 0.033 0.08 Intensity at Percent Yield Intensity Intensity Intensity at 400nm, 483nm Area (400- of xylose at 400nm at 483nm 527nm and 527nm 650nm) 3hr 0.139 0.008 0.001 0.1988 0.079 24hr 0.166 0.0018 0.001 0.31 0.105 48hr 0.2155 0.01 17 0.0009 0.326 0.0967 Principal Component Regression (PCR) between fluorescence data and sugar values PCA (Principal Component Analysis) deals with data sets with high dimensionality. PCA is used in statistics to reduce the dimensionality of large data sets and identity new significant variables. One way to find the Principal Components of a data set is by calculating the eigenvectors of the data correlation matrix. These vectors give the directions in which the data cloud is most stretched. The projections of the data on the eigenvectors are the Principal Components. The corresponding eigenvalues indicate the amount of information the respective Principal Components represent. Principal Components corresponding to large eigenvalues account for most of the variability and thus tell us much about the association between the data points. After analysis is done on the data, multiple linear regression (MLR) is done between the 97 principal components chosen from the fluorescence data and the concentration or percent yield of sugar. The r? for the PCR analysis between the fluorescence data and the sugar values are given in Table 5.4. Table 5.4 R2 values from PCR analysis of fluorescence data and sugar r2 values dextrose xylose total dextrose xylose L cane. cane. cane. yield yield otal yield 3hr 0.232 0.216 0.231 0.17 0.149 0.145 24hr 0.369 0.08 0.262 0.37 0.175 0.34 48hr 0.232 0.056 0.129 0.333 0.186 0.3 We conclude that there is not much correlation between the chosen principal components of the fluorescence data and the concentration or percent yield of sugar. The largest 12 values were obtained for the regression between dextrose at 24hr and the fluorescence spectra but the 1*2 value was only 0.37. Since both Cd and lignin content effect digestibility of biomass, we combined the data from the Cd and the principal components from the fluorescence data. The combined data was then regressed against the sugar concentration and percent yield. Table 5.2 gives the r2 values for the correlation studies and the highest 1'2 value was obtained for correlation between the combined data and the concentration of dextrose at 24hr. Table 5.5 R2 values from PCR analysis of combination of fluorescence and Crl data versus sugar r2 values dextrose xylose total dextrose xylose total conc. cone. conc. yield yield yield 3hr 0.45 0.38 0.45 0.38 0.34 0.366 24hr 0.515 0.38 0.47 0.48 0.44 0.46 48hr 0.48 0.45 0.48 0.44 0.53 0.47 98 One of the main reasons that no correlation was found between the fluorescence data and the sugar values by using either multiple linear regression or principal component regression was because we could not verify that the fluorescence method was measuring lignin content. The delignified rice straw sample fluoresced even though it did not contain any lignin. Although Billa et al. (1999) found a good correlation between fluorescence data and lignin content; the statistical analysis may have been flawed. By repeating the experiments at 450, 400, 350 and 280nm and incorporating all the data points into the x-matrix they are being redundant because the data at one wavelength may correlate highly with the data measured at another wavelength. By adding redundant data, one can get very high correlations but this method of statistical analysis can give high r2 values but the correlation is not true. Also, the correlations are not extremely high because the highest correlation they get is an 1‘2 value of 0.77. Furthermore, they remove some outliers to increase the r2 value to 0.86. Another reason we did not get any correlations is because we did not extract the lignin content of the rice straw as Billa et al. (1999) did. They correlated extracted lignin content with the fluorescence data and found moderately good correlations. We correlated the sugar concentration and yield to fluorescence data assuming that lignin content has the maximum effect on digestibility of lignocellulosics such as rice straw but as seen in the hydrolysis of delignified straw, lignin has only a moderate effect on the digestibility of rice straw. Also, rice straw is composed of 14% Klason lignin and 15.5% ash (Wei and Cheng, 1985). Maybe the treatments are breaking the lignocellulosic bonds and degrading the li gnin polymer but even the li gnin monomers will fluorescence since 99 they are also composed of phenolic groups. Therefore, fluorescence is not a good technique to predict the digestibility of AFEX treated rice straw. Multiple Linear Regression between fluorescence data and Treatment Conditions The initial treatment conditions of the AFEX process were also regressed against each of the sugar values. The results are shown in Table 5.6. Table 5.6 R2 Values from MLR Analysis of Fluorescence Data and Treatment Conditions ’2 Area Intensity Intensity Intensity (400- at 400nm at 483nm at 527nm 650nm) Treatment ponditons 0.285 0.355 0.515 0.0486 No correlation was found between the initial conditions and the fluorescence data. The highest 12 was 0.515 for the peak at 527nm but an r2 of 0.515 is not significant due to high scatter. As we mentioned earlier, the lignin monomers will also fluoresce even if the treatment is affecting the lignin polymer therefore, there is no way to predict how the treatment will affect the composition of lignin within the rice straw matrix. 5.4 Summary Lignin sterically hinders the cellulase enzymes from coming in contact with the crystalline cellulose thus inhibiting enzymatic hydrolysis. AFEX treatment was proposed to have disrupted the lignin polymers within the rice straw and broken the lignin into fragments. The proposed hypothesis could not be proved using fluorescence. AFEX treatment does not seem to affect the fluorescence profile of the various rice straw samples. However, we may not be getting an accurate picture of the change in the lignin 100 content of the rice straw. The problem using the fluorescence technique is that even the broken lignin monomers will fluoresce and as seen by the fluorescence profile of the deli gnified rice straw, rice straw without any lignin also fluoresces. Finally, as mentioned earlier, crystallinity of cellulose also plays an important role in the rate of enzymatic hydrolysis. Both crystallinity and lignin content affect the hydrolysis. Thus, we cannot use the fluorescence technique alone to predict digestibility of rice straw. Combining the Cd and fluorescence data, we obtained only an 1’2 value of 0.5 and therefore, both Cd and fluorescence data could not predict the digestibility of rice straw. 101 6 DRIFT 6.1 Introduction Bonds between atoms are always in motion and the molecule when irradiated with a certain wavelength of radiation energy absorbs energy. The different molecular stretching and bending motions in the infrared region are; symmetric, antisymmetric, in plane bending and out of plane bending. Infrared absorptions result from changes in the vibrational and rotational state of a molecular bond coupling with electromagnetic radiation if the vibrating molecule produces an oscillating dipole moment that can interact with the electric field of radiation. The amount of energy absorbed depends on the size of the atoms involved and the length of the bond. Thus Infrared (IR) spectrum gives us information on the types of bonds or the functional groups in a molecule. Diffuse Reflectance Infrared Fourier Transform (DRIFT) spectrometry is an IR technique useful for analysis of powders and rough surface solids. In diffuse reflectance, the incident light is scattered from the surface of the sample and is collected by a parabolic mirror and passed to a detector. In DRIFT applications, the radiation reflected from the samples is due to both diffuse reflectance and specular reflectance. Specular reflectance is defined as IR radiation reflected from the sample surface, which does not penetrate the sample. Multiple vibrations occurring simultaneously create a highly complex absorption spectrum uniquely characteristic of the functional groups that make up the molecule. DRIFT methods have previously contributed to the characterization of the chemical components of wood, pulp and straw. However, since lignocellulosic materials are highly complex, most of the FTIR observed bands couldn’t be directly assigned to one single component (Owen and Thomas, 1989) but using multivariate analyses Ferraz et al. 102 (2000) was able to predict the concentration of important wood components (glucans, xylans, lignin) with r2 values higher than 0.86. Therefore, we used the DRIFT method to predict digestibility of AFEX treated rice straw materials by correlating spectral data with the glucose and xylose sugar concentration and yields. 6.2 Experimental 6.2.1 DRIFT Spectroscopy Method A Perkin Elmer System 2000 FT-IR with the Diffuse Reflectance Infrared Fourier Transform Spectrometry (DRIFT ) accessory was used for DRIFT analysis of rice straw. About 25 mg of the 106 um sieved rice straw was homogenized with 250 mg of KBr to a final concentration of 10%. Infrared Radiation (IR) reflected from the sample surface without penetration of radiation is known as specular reflectance. Specular reflectance produces inverted bands (restsrahlen bands), which should be minimized or eliminated by grinding the sample and diluting with KBr. Duplicate spectra were recorded from 4000 to 800 cm“1 using 64 scans, triangular apodization and resolution of 4 cm]. The data points were set apart by 1 cm'1 and the reflectance spectra were transformed to Kubelka- Munk (KM) units to minimize scattering contributions to the absorption measured (Backa and Brolin, 1991). Total analysis time is lO-lSmin/sample excluding sample preparations. Conversion of spectra to Kubelka-Munk: Percent Transmittance = 10" (2 - absorbance) (6.1) ((1 — Transmittarce)‘2) 2 * Transmittarce Kubelka- Munk = (6.2) 103 6.2.2 Principal Component Regression DRIFT spectra were analyzed on MatLab. The spectra were normalized to 1507cm'l and then mean scaled (covariance about the mean) by subtracting each spectrum from the mean spectrum. The properties of the rice straw were indicated by the data of DRIFT spectra. The same algorithm for PCA analysis of fluorescence spectra was also used for PCA analysis of DRIFT spectra. Principal component regression (PCR) models were based on Principal Component Analysis (PCA) performed on spectral information followed by Multiple Linear Regression (MIR) between sugar concentrations and principal components. The regression coefficients were set at 99% confidence interval. The analysis of variance (ANOVA) for each model gave the r2 values that were used to validate the models. 6.3 Results and Discussion 6.3.1 Spectroscopic Characterization of AFEX Treated Rice Straw The samples were analyzed with rice straw sieved through ISO-mesh (106 um openings), because particle size has an effect on the light scattering. According to Backa and Brolin (1991), the precision of DRIFT analysis increases for samples less than 160 pm. No other particle size was analyzed because sieving it through an opening of 106 um gave powdered rice straw. A non-linear relationship exists between concentration and absorbency in Kubelka-Munk units when the sample to dilutent ratio exceeds 0.3 due to saturation effects because the penetration of the incident light becomes independent of absorptivity. Therefore, we diluted our samples to 10% with KBr. We thoroughly homogenized the samples with KBr because it improves precision of the DRIFT spectra (Backa and Brolin, 1991). 104 All AFEX treated rice straw samples were analyzed by DRIFT spectroscopy. The pure band at 1510 cm'1 is related to the aromatic components present in lignin and is usually used as the reference band (Pandey, 1999). This band is purely due to the aromatic skeletal vibration of benzene ring in lignin. In our spectra, the internal reference band shifted a few degrees from 1510 cm'1 to 1507". Therefore, the spectra of the samples were normalized to the absorption intensity at 1507 cm". Reproducibility To study if the error in the DRIFT spectra came from a new sample or it was due to the way sample was placed in the sample holder, we obtained spectra of replaced sample and spectra of newly ground sample. Spectra for replicate samples of BI] 1 as illustrated in Figure 6.1 provided adequate reproducibility. Therefore, the spectra were independent of the way the sample was placed in the holder and of grinding. Also, the inherent composition of the rice straw did not vary as seen by the reproducibility plot. 105 Intensity (Rubella-Munk) 3 IfTI I I I I l T I I I I I Wfirj I I I I I T r I I I I I I I I l I I I IJ I— 8111 ' -°- newsample b . l- o 500 10001500200025003000350040004500 Wavenumber (6'5") Figure 6.1 Replicate DRIFT spectra of B111 (Samples were checked for reproducibility) 106 However, only the bands in the region 2987-3321 cm’1 and 1038-1209 cm'1 lacked reproducibility. Distortions in the region from 1038-1209 cm'1 are attributed to the specular radiation while the distortions in the region from 2987-3483 cm'1 are attributed to the O-H stretching (Pandey, 1999). Probably the error seen in the region 2987- 3482cm'l is due to the different amounts of moisture in the KBr and the straw. The intensity of the bands at 1660 cm'1 is from the conjugated (C=O) bond stretching while the band at 1750cm "' is from the stretching of the unconjugated (C=O) bond. Other bands such as the O-H stretch at around 3400 cm’l, C-H stretch at 2900 cm"1 and C=C stretch at around 1510 cm'1 are also pure bands while bands below 1460cm-1 are complex, having contributions from various vibration modes in carbohydrates and lignin (Faix et al., 1992; Pandey and Theagarajan, 1997; Harrington et al., 1964). Effect of Delignification 0n DRIFT Spectra Delignification of rice straw had a tremendous effect on the DRIFT spectra as seen in Figure 6.2. Even a new band around 2300cm-l is formed which is difficult to attribute to a particular species since many molecules show absorption bands in the region. Bands at 2987-3483cm-l and 1038-1209cm-1 were much lower due to reduced specular radiation and lower absorbed water. The band at 1300-1800cm-l is really affected by delignification. This region is the fingerprint region of complex carbohydrates and 1i gnin. The permanganate oxidation seems to have oxidized the lignin as well as other aromatic and unsaturated molecules in the rice straw as seen in previous research (Edwards and Mackney, 1939). 107 Effect of Delignification 4 I. I j—r I I I I Ifi I fir I I T Z I _._ Delignified Rice Straw 3 _ i. - - B - - Untreated Rice Straw " I Intensity (Kubelka-Munk) N l O I I r I I I I I I 5001000150020002500300035004000 Wavenumber (cm '1) Figure 6.2 Effect of Delignification on the DRIFT Spectra of Rice Straw 108 Efiect of AFEX treatment on DRIFT Spectra The effect of the AFEX treatment on the DRIFT spectra is shown in Figure 6.3 and Figure 6.4. A spectral comparison of samples with the same sugar yields along with untreated rice straw is shown in Figure 6.3. In this figure the spectra of B1 11 and 3126 overlap but the rice straw varies slightly from both spectra. Untreated rice straw varies mostly in the region from 2300 to 3200 cm". The region of specular radiation and water absorption region was not included for the comparisons. In Figure 6.4, we compared spectra of two samples with different yields of sugars. The samples differed in the fingerprint region of 1400-1800cm'l for complex carbohydrates and lignin. The peak around 1740 cm-1 is ascribed to holocellulose (cellulose and hemicellulose) because carbonyl groups are found mostly in branched chain hemicelluloses (Pandey, 1999). Maybe the AFEX process is increasing the yields of xylose by affecting the hemicelluloses within the rice straw. AFEX treatrrrent did affect the DRIFT spectra in this fingerprint region although the effect was not dramatic. 109 a. .I I ITI—FIITjfiTIIII I to 01 ....ar t 1 . - — B111 ------- Untreated Rice Straw -—a— 8126 —' 00 IIIIIIIIIIIIIIIIIIIIIIIIIIIII P or llJLlIllll Intenslty (Kubelka-Munk) N 1.5 1 1 .J 0.5 ;- __ C I 0 r1 1 1 LLJ r r r I i r r r r r I r r r r l r r r r I r 4 500 1000 1500 2000.2500 3000 3500 4000 4500 Wavenumber (cm -1) Figure 6.3 Effect of Various AFEX treatments on DRIFT spectra B111 (60 % moisture, 80 °C, 10 min, ammonia ratio of l) B126 (20 % moisture, 80 °C, 10 min, ammonia ratio of 2) 110 +8150 “#8111 A 3 E 1.5 — i Intenslty (Kubelka-Munk) q 05- r o b -1 >- q 1— . .1 I- J o I r i x l r J_A ' 14;; 4m 1 l r 14 l I 1_L r r I l 1 A A I 1 - L———' r J r r 500 1000 1500 2000 2500 3000 3500 4000 4500 Wavenumbers (cm -‘) Figure 6.4 Comparion of Spectra with different Sugar Yields at 48hr Enzymatic Hydrolysis B111 (38% Dextrose, 30% Xylose) B150 (81% Dextrose, 80% Xylose) 111 6.3.2 Multivariate Analysis Since the analysis of IR spectra is complicated, the use of multivariate analysis allows for the simultaneous analysis of the entire spectral data. The entire spectral range from 400-800cm'l was regressed using PCA algorithm. PCR Analysis After PCA analysis on the spectral data, only 10 principal components were chosen for MLR analysis between the sugar values and the chosen principal components. The results of the PCR analysis between DRIFT spectral data and the concentration or percent yield of sugar are given in Table 6.1. 112 Table 6.1 R2 values for PCR analysis between DRIFT spectra and sugar r2 values I dextrose xylose total dextrose xylose L J cone. cane. cane. yield yield otal yiel 3hr 0.39 0.303 0.343 0.5 0.255 0.41 | 24hr 0.29 0.288 0.294 0.347 0.23 0.34 | 48hr 0.234 0.2 0.2 0.426 0.223 0.35 I The highest r2 value of 0.5 was obtained from MLR between percent yield of dextrose at 3hr and the chosen principal components of the DRIFT spectra. Only 50 percent of the error was explained from the PLR analysis between the percent yield of dextrose at 3hr and DRIFT spectra. An r2 value of 0.5 allows for a lot of scatter in the data set and therefore, the percent yield of dextrose at 3hr is not a very good predictor of digestibility of AFEX treated rice straw. Most probably the correlation would be much better if the error in dextrose yield was lower. Assuming that the r2 value of 0.5 would increase with better hydrolysis results, we can say that probably the lignin content is affecting the initial hydrolysis rates of cellulose. If the enzymes can get past the lignin barrier, the CrI of the biomass then determines the ultimate yield of dextrose. Therefore, initial hydrolysis rates are determined by how much enzyme is getting past the lignin hindrance. Both lignin and Cd govern ultimate yields of sugar. Since Cd and lignin content have the greatest effect on the hydrolysis of biomass, we combined both the Cd and DRIFT principal components to predict the digestibility of AFEX treated rice straw samples. The correlation between the combined data slightly increased the r2 values as seen in Table 6.2. Again the highest 1'2 value was seen for the correlation between the combined data and 3hr yield of dextrose. 113 Table 6.2 R2 values for PCR analysis between combination of DRIFT and Crl data and sugar I? values dextrose xylose total dextrose xylose total cone. cone. cone. yield yield yield 3hr 0.42 0.34 0.39 0.53 0.3 0.47 24hr 0.33 0.37 0.31 0.33 0.4 0.34 48hr 0.32 0.35 0.34 0.43 0.42 0.44 Ferraz et al. (2000) correlated extracted lignin content with the DRIFT data. We skipped steps by directly correlating DRIFT data with sugar values from the various AFEX treated rice straw samples. Therefore, the correlation did not give good results because either lignin content is not a good predictor of digestibility or maybe because lignin content is not varying much with treatment. 6.4 Summary The digestibility of the AFEX treated rice straw could not be predicted using the DRIFT spectra. An r2 value of 0.5 was obtained for PCR analysis between 3hr dextrose yield and DRIFT data. Combination of the Crl and DRIFT principal components gave a slightly higher r2 value of 0.53. There was not a striking difference among the DRIFT spectra of various samples and no new bands are observed among the spectra. As mentioned earlier, either 1i gnin content is not a good predictor of digestibility or the lignin content is not varying much within the rice straw. Also, there might not be a difference between 1i gnin monomers and the polymers as seen in the DRIFT spectra. Crystallinity of the cellulose also affects the digestibility of the rice straw. Therefore, lignin content alone cannot be used to predict digestibility of the rice straw and also the combination of Cd data and the DRIFT principal components could not be used to predict the digestibility of the rice straw. 114 7 CONCLUSION Our research goal was to predict the digestibility of the AFEX treated rice straw. We proposed that lignin and Crl of the rice straw were the main factors, which determine the digestibility of the rice straw and that one of these factors could be used to predict digestibility of AFEX treated rice straw. We tried to model the digestibility of the rice straw by statistically correlating the variability of the samples to differences in treatment using X-ray diffraction, fluorescence spectroscopy and DRIFT spectroscopy. Crystallinity index (CR1) was calculated using x-ray diffraction while 1i gnin content of the samples was predicted using fluorescence and DRIFT spectroscopy. Multi-variant analysis of the sugar values versus the spectral data did not give us good correlations. None of the analytical techniques were good predictors of the digestibility of the AFEX treated rice straw samples. Not much variation was found in the DRIFT and fluorescence spectra of the samples treated at different process conditions. This is probably because even the disrupted 1i gnin monomers fluoresce and have similar rotational spectroscopy as the lignin polymers. Some interesting facts to note were that Crl decreased for most of the AFEX treated samples except for about three samples. These samples with a higher CrI still had greater yields of sugar than untreated rice straw. Therefore, we concluded that Crl of some samples was increasing due to removal of amorphous lignin, cellulose and hemicellulose (Gharpuray et al., 1983). Also, fluorescence spectroscopy could not prove the presence of 1i gnin because delignified rice straw also fluoresced with the same intensity as untreated rice straw. Among the various analytical techniques, DRIFT spectroscopy gave the best results with an 1'2 value of 0.5 for PCR analysis between 3 hr dextrose yield and the chosen principal 115 components. Since DRIFT spectroscopy has been able to predict the concentration of lignin (Ferraz et al., 2000), perhaps lignin content determines the initial rate of hydrolysis. Hindrance of enzymes of lignin from adsorbing to the substrate may determine the initial rate of hydrolysis. CrI then determines the amount of substrate, which will be digested by the enzymes. The greater the CrI the slower is the hydrolysis rate and the opposite is true also. Both lignin content and Cd probably determine ultimate yields (Chang and Holtzapple, 2000). The best correlation was found between the initial treatment conditions versus the concentration of xylose at 24 hr enzymatic hydrolysis. We obtained an r2 of 0.6 for this correlation. Since hemicellulose is composed primarily of xylose units, it seems as though the initial treatments are directly affecting hemicellulose content by solubilization of hemicellulose or removal of acetyl groups (Bouveng,]96l). On the contrary, the hydrolysis of cellulose is dependent on many factors such as lignin content and crystallinity of cellulose. The principal problem with our statistical analysis was from the reproducibility data of sugar. We had 30% error in the concentration of sugar and another 10% error with obtaining the potential amount of sugar from acid hydrolysis. We propose that the correlations will be greater for more reproducible sugar yields. 116 7.1 Recommendations 1. Improvement of AFEX process so that we can have reproducible hydrolysis results. Enzymatic hydrolysis experiments should be tried at a larger scale. Determine sugar concentrations on the DIONEX system to obtain more accurate concentration of sugars. DRIFT analysis of the more improved AFEX process samples. Multi-variant analysis using Partial Least Square Regression analysis improves the statistical analysis slightly compared to Principal Component Analysis. 117 8 APPENDIX 118 Table 8.1 Concentration of Sugar at 3hr after Enzymatic Digestion Initial conditions 3hr sugar cane. Treatment Treament Treatment Treatment ID Moisture Time Temperature Ammonia dim“ x231? (%)* (min) (°C) Ratio (*‘) 8103 60 1 0 90 0.5 0.1 1 0.06 8105 60 10 90 0.75 0.14 0.10 8108 60 10 90 1 0.10 0.07 811 20 5 80 1 0.18 0.07 8111 60 10 80 1 0.11 0.08 8114 20 5 80 1.5 0.29 0.17 8120 20 5 90 1.5 0.20 0.08 8123 20 10 80 1.5 0.15 0.11 8126 20 10 80 2 0.17 0.09 8135 20 10 90 1.5 0.23 0.16 8138 20 10 90 2 0.29 0.14 814 20 10 80 0.5 0.15 0.06 8141 40 5 80 1.5 0.10 0.08 8144 40 5 80 2 0.09 0.05 8147 40 5 90 1.5 0.23 0.17 8150 40 5 90 2 0.23 0.14 8153 40 10 80 1.5 0.17 0.11 8156 40 10 90 1.5 0.12 0.14 8159 40 10 90 2 0.17 0.15 8162 40 10 80 2 0.13 0.10 8165 60 5 80 1.5 0.11 0.08 8168 60 5 80 2 0.10 0.07 817 20 10 80 0.75 0.21 0.08 8177 60 5 90 1.5 0.17 0.13 * Treatment Moisture ((Kg water/Kg dry straw)x100) ** Treatment Ammonia Ratio (lb ammonia/lb dry straw) 119 Table 8.1 (Cont’d) Initial conditions 3hr sugar cane. Treatment Treatment Treament ID Moisture :mtmfi; Temperature Ammonia dist/3547,5739 (%)* (°C) Ratio (“) B180 60 5 90 2 0.09 0.05 8186 60 10 90 2 0.15 0.13 82 20 5 0.5 0.16 0.05 820 20 10 90 0.5 0.10 0.05 824 20 10 90 0.75 0.08 0.06 828 20 10 90 1 0.17 0.14 832 20 5 90 0.1 0.19 0.07 835 20 5 90 0.75 0.14 0.07 838 20 5 90 1 0.06 0.05 841 20 10 80 1 0.25 0.17 845 40 5 80 0.75 0.19 0.07 85 40 5 0.5 0.16 0.05 851 40 5 90 0.5 0.19 0.06 854 40 5 90 0.75 0.10 0.06 857 40 10 80 0.5 0.13 0.08 860 40 10 80 0.75 0.13 0.09 863 40 5 90 1 0.09 0.07 866 40 10 80 1 0.11 0.10 869 40 10 90 0.5 0.14 0.05 872 40 10 90 0.75 0.22 0.11 875 40 10 90 0.1 0.10 0.10 878 60 5 80 0.5 0.08 0.05 88 20 5 80 0.75 0.14 0.05 884 60 5 80 1 0.20 0.06 887 60 5 90 0.5 0.09 0.04 890 60 5 90 0.75 0.14 0.07 C172 60 10 80 1.5 0.07 0.07 C184 60 10 90 1.5 0.13 0.12 C49 40 5 80 1 0.15 0.09 untreated rice straw 0.10 0.03 * Treatment Moisture ((Kg water/Kg dry straw) x100) ** Treatment Ammonia Ratio (lb ammonia/lb dry straw) 120 Table 8.2 Concentration of Sugar at 24hr Enzymatic Digestion Initial conditions 24hr sugar cone. Treatment Treatment Treatment ID Moisture :mtmmt) Temperature Ammonia “(Stiff)“ x375: (%)* (°C) Ratio (“) B103 60 10 90 0.5 0.36 0.12 8104 60 10 80 2 0.33 0.2 8105 60 10 90 0.75 0.45 0.17 8108 60 10 90 1 0.22 0.11 811 20 5 80 1. 0.32 0.09 8111 60 10 80 1 0.28 0.12 8114 20 5 80 1.5 0.45 0.2 8120 20 5 90 1.5 0.43 0.15 8123 20 10 80 1.5 0.53 0.28 8126 20 10 80 2 0.64 0.28 8135 20 10 90 1.5 0.52 0.27 8138 20 10 90 2 0.48 0.22 814 20 10 80 0.5 0.38 0.11 8141 40 5 80 1.5 0.28 0.2 8144 40 5 80 2 0.36 0.19 8147 40 5 90 1.5 0.49 0.28 8150 40 5 90 2 0.53 0.29 8153 40 10 80 1.5 0.41 0.31 8156 40 10 90 1.5 0.36 0.36 8159 40 10 90 2 0.35 0.32 8162 40 10 80 2 0.35 0.24 8165 60 5 80 1.5 0.36 0.17 8168 60 5 80 2 022 0.13 817 20 10 80 0.75 0.41 0.12 8177 60 5 90 1.5 0.38 0.24 * Treatment Moisture ((Kg water/Kg dry straw) x100) ** Treatment Ammonia Ratio (lb ammonia/lb dry straw) 121 Table 8.2 (Cont’d) 24hr Initial conditions sugar cane. Treatment Wreatmen Treatment Treatment ID Moisture Time 1Temperature Ammonia “‘33“ 7;? (%)" (min) (°C) ratio (”) 81 80 60 5 90 2 0.21 0.07 8186 60 1 0 90 2 0.46 0.35 82 20 5 0.5 0.31 0.09 820 20 10 90 0.5 0.3 0.08 824 20 1 0 90 0.75 0.28 0.13 828 20 10 90 1 0.36 0.22 832 20 5 90 0.1 0.42 0.13 835 20 5 90 0.75 0.22 0.08 838 20 5 90 1 0.19 0.08 841 20 1 0 80 1 0.36 0.19 845 40 5 80 0.75 0.38 0.1 1 BS 40 5 0.5 0.44 0.09 851 40 5 90 0.5 0.36 0.1 854 40 5 90 0.75 0.36 0.1 1 BS7 40 1 0 80 0.5 0.49 0.17 B60 40 10 80 0.75 0.35 0.14 863 40 5 90 1 0.33 0.16 866 40 10 80 1 0.32 0.21 869 40 10 90 0.5 0.41 0.1 1 872 40 1 0 90 0.75 0.38 0.16 875 40 10 90 0.1 0.33 0.18 878 60 5 80 0.5 0.35 0.09 88 20 5 80 0.75 0.34 0.1 884 60 5 80 1 0.42 0.1 3 887 60 5 90 0.5 0.21 0.06 890 60 5 90 0.75 0.36 0.1 0172 60 10 80 1 .5 0.25 0.23 C184 60 10 90 1 .5 0.24 0.22 C49 40 5 80 1 0.4 0.1 6 untreated rice 0.17 0.04 straw * Treatment Moisture ((Kg water/Kg dry straw) x 100) ** Treatment Ammonia Ratio (lb ammonia/lb dry 122 Table 8.3 Concentration of Sugar at 48hr Enzymatic Digestion 48hr Initial conditions sugar cane. Treatment Treatment Treatment ID Moisture :mtmzt) Temperature Ammonia dim” 7339 (°/.)* (°C) Ratio (“) 8103 60 10 90 0.5 0.47 0.15 8104 60 10 80 2 0.52 0.29 8105 60 10 90 0.75 0.54 0.21 8108 60 10 90 1 0.37 0.17 811 20 5 80 1 0.39 0.11 8111 60 10 80 1 0.44 0.18 8114 20 5 80 1.5 0.55 0.25 8120 20 5 90 1.5 0.59 0.21 8123 20 10 80 1.5 0.6 0.31 8126 20 10 80 2 0.56 0.26 8135 20 10 90 1.5 0.66 0.33 8138 20 10 90 2 0.61 0.27 814 20 10 80 0.5 0.5 0.2 8141 40 5 80 1.5 0.48 0.35 8144 40 5 80 2 0.48 0.25 8147 40 5 90 1.5 0.73 0.38 8150 40 5 90 2 0.73 0.36 8153 40 10 80 1.5 0.63 0.41 8156 40 10 90 1.5 0.57 0.47 8159 40 10 90 2 0.54 0.41 8162 40 10 80 2 0.57 0.36 8165 60 5 80 1.5 0.37 0.17 8168 60 5 80 2 0.4 0.23 817 20 10 80 0.75 0.5 0.16 8177 60 5 90 1.5 0.6 0.33 * Treatment Moisture ((Kg water/Kg dry straw) x100) ** Treatment Ammonia Ratio (lb ammonia/lb dry 123 Table 8.3 (Cont’d) 48hr Initial conditions sugar cone. Treatment Treatment Treatment ID Moisture :5;ng Temperature Ammonia d722,)“ x221? (%)* (°C) Ratio (“) B180 60 5 90 2 0.36 0.11 8186 60 10 90 2 0.58 0.4 82 20 5 0.5 0.36 0.09 820 20 10 90 0.5 0.42 0.1 1 824 20 10 90 0.75 - 0.45 0.19 828 20 10 90 1 0.47 0.26 832 20 5 90 0.1 0.53 0.17 835 20 5 90 0.75 0.32 0.1 1 B38 20 5 90 1 0.37 0.13 841 20 10 80 1 0.47 0.23 845 40 5 80 0.75 0.5 0.16 85 40 5 0.5 0.5 0.1 851 40 5 90 0.5 0.42 0.1 1 854 40 5 90 0.75 0.45 0.14 857 40 10 80 0.5 0.54 0.19 860 40 10 80 0.75 0.17 0.18 863 40 5 90 1 0.39 0.22 866 40 10 80 1 0.49 0.29 869 40 10 90 0.5 0.45 0.13 872 40 10 90 0.75 0.46 0.17 875 40 10 90 0.1 0.44 0.23 878 60 5 80 0.5 0.47 0.12 88 20 5 80 0.75 0.47 0.14 884 60 5 80 1 0.54 0.17 890 60 5 90 0.75 0.44 0.13 C172 60 10 80 1.5 0.46 0.37 C184 60 10 90 1.5 0.4 0.38 049 40 5 80 1 0.48 0.16 Untreated rice straw 0.22 0.06 * Treatment Moisture ((Kg water/Kg dry straw) x100) ** Treatment Ammonia Ratio (lb ammonia/lb dry 124 Table 8.4 Percent Yield of Sugar at 3hr Enzymatic Digestion Initial conditions Treatment Treatment Treatment Treatment Yield of Yield of ID Moisture Time Temperature: Ammonia dextrose xylose (%) (%)* (min) (°C) Ratio (“) (%) 8103 60 10 90 0.5 7.79 7.39 8105 60 10 90 0.75 16.47 20.00 8108 60 10 90 1 7.86 11.49 811 20 5 80 1 16.36 11.67 8111 60 10 80 1 9.57 13.33 8114 20 5 80 1.5 29.00 30.91 8120 20 5 90 1.5 16.00 11.43 8123 20 10 80 1.5 13.04 18.33 8126 20 10 80 2 11.72 12.00 8135 20 10 90 1.5 18.40 22.86 8138 20 10 90 2 23.20 20.00 814 20 10 80 0.5 15.00 12.00 8141 40 5 80 1.5 7.92 11.48 8144 40 5 80 2 7.87 7.92 8147 40 5 90 1.5 17.04 22.67 8150 40 5 90 2 25.56 31.11 8153 40 10 80 1.5 12.21 14.64 8156 40 10 90 1.5 8.75 18.27 8159 40 10 90 2 12.82 19.75 8162 40 10 80 2 9.61 13.99 8165 60 5 80 1 .5 8.80 1 1 .43 B168 60 5 80 2 10.18 12.42 817 20 10 80 0.75 28.00 20.00 8177 60 5 90 1.5 20.00 28.89 * Treatment Moisture ((Kg water/Kg dry straw) x 100) ** Treatment Ammonia Ratio (lb ammonia/lb dry straw) 125 Table 8.4 (Cont’d) Initial conditions Treatment Treatment Treatment Yield of ID Moisture 11mm?) Temperature Ammonia dextrose “VIE: 3;.) (%i‘ (°C) Ratio ("i (%) 8180 60 5 90 2 14.58 15.34 81 86 60 10 90 2 12.72 18.56 820 20 10 90 0.5 9.49 8.98 824 20 10 90 0.75 10.81 16.18 828 20 10 90 1 14.78 23.33 832 20 5 90 0.1 17.27 11.67 835 20 5 90 0.75 12.73 11.67 841 20 10 80 1 17.86 26.15 845 40 5 80 0.75 27.14 17.50 851 40 5 90 0.5 27.14 15.00 854 40 5 90 0.75 8.70 8.57 857 40 10 80 0.5 1 1 .30 12.31 860 40 10 80 0.75 11.82 15.00 863 40 5 90 1 7.83 1 1.23 866 40 10 80 1 9.23 16.68 869 40 10 90 0.5 1 1.20 6.67 872 40 10 90 0.75 19.13 18.33 875 40 10 90 0.1 13.33 28.57 878 60 5 80 0.5 12.31 14.29 88 20 5 80 0.75 12.17 7.69 884 60 5 80 1 19.05 10.00 887 60 5 90 0.5 6.92 5.13 890 60 5 90 0.75 10.77 9.33 C172 60 10 80 1.5 7.07 13.98 C184 60 10 90 1.5 10.66 18.60 C49 40 5 80 1 13.64 15.00 untreated rice straw 10.17 5.44 * Treatment Moisture ((Kg water/Kg dry straw) x 100) ** Treatment Ammonia Ratio (lb ammonia/lb dry 126 Table 8.5 Percent Yield of Sugar at 24hr after Enzymatic Digestion Initial conditions Treatment Treatment Treatment Treatment Yield of Yield ct ID Moisture Time (min) Temperature Ammonia dextrose xylose (%)* (°C) Ratio (“) (%) (%) B103 60 10 90 0.5 26.67 16.00 8104 60 10 80 2 22.00 25.00 8105 60 10 90 0.75 52.94 34.00 81 08 60 10 90 1 17.60 16.92 811 20 5 80 1 29.09 15.00 81 1 1 60 10 80 1 24.35 20.00 8114 20 5 80 1.5 45.00 36.36 8120 20 5 90 1 .5 34.40 21.43 81 23 20 10 80 1 .5 46.09 46.67 8126 20 10 80 2 44.14 37.33 8135 20 10 90 1.5 41.60 38.57 81 38 20 10 90 2 38.40 31 .43 B14 20 10 80 0.5 38.00 22.00 8141 40 5 80 1 .5 22.40 28.37 8144 40 5 80 2 32.73 31.67 8147 40 5 90 '1 .5 36.30 37.33 8150 40 5 90 2 58.89 64.44 8153 40 10 80 1.5 29.29 41.33 81 56 40 10 90 1 .5 26.67 48.00 8159 40 10 90 2 26.92 42.67 8162 40 10 80 2 25.93 32.00 8165 60 5 80 1 .5 28.80 24.29 8168 60 5 80 2 23.16 23.64 817 20 10 80 0.75 54.67 30.00 8177 60 5 90 1.5 44.71 53.33 * Treatment Moisture ((Kg water/Kg dry straw) x 100) ** Treatment Ammonia Ratio (1b ammonia/lb dry straw) 127 Table 8.5 (Cont’d) Initial conditions Treatment Treatment Treatment Treatment Yield of Yield of ID Moisture Time (min) Temperature Ammonia dextrose xylose (%I‘ (°C) ratio (%) (%) B180 60 5 90 2 32.31 20.00 8186 60 10 90 2 38.33 50.00 820 20 10 90 0.5 27.27 13.33 824 20 10 90 0.75 37.33 32.50 828 20 10 90 1 31.30 36.67 832 20 5 90 0.1 38.18 21.67 835 20 5 90 0.75 20.00 13.33 841 20 10 80 1 25.71 29.23 845 40 5 80 0.75 54.29 27.50 851 40 5 90 0.5 51 .43 25.00 854 40 5 90 0.75 31.30 15.71 857 40 10 80 0.5 42.61 26.15 860 40 10 80 0.75 31.82 23.33 863 40 5 90 1 28.70 26.67 866 40 10 80 1 27.83 35.00 869 40 10 90 0.5 32.80 14.67 872 40 10 90 0.75 33.04 26.67 875 40 10 90 0.1 44.00 51.43 878 60 5 80 0.5 53.85 25.71 88 20 5 80 0.75 29.57 15.38 884 60 5 80 1 40.00 21.67 887 60 5 90 0.5 16.80 8.57 890 60 5 90 0.75 27.69 13.33 C172 60 10 80 1.5 25.00 46.00 C184 60 10 90 1.5 19.20 33.85 C49 40 5 80 1 36.36 26.67 untreated rice straw 17.00 8.00 * Treatment Moisture ((Kg water/Kg dry straw) x 100) ** Treatment Ammonia Ratio (lb ammonia/lb dry 128 J_tll. . Table 8.6 Percent Yield of Sugar at 48hr Enzymatic Digestion Initial conditions Treatment Treatment Treatment Treatment Yield of Yield of ID Moisture Time Temperature Ammonia dextrose xylose (%)* (min) (°C) ratio (“1 (%) (%) B103 60 10 90 0.5 34.81 20.00 8104 60 10 80 2 34.67 36.25 8105 60 10 90 0.75 63.53 42.00 8108 60 10 90 1 29.60 26.15 811 20 5 80 1 35.45 18.33 8111 60 10 80 1 38.26 30.00 8114 20 5 80 1.5 55.00 45.45 8120 20 5 90 1.5 47.20 30.00 8123 20 10 80 1.5 55.65 51.67 8126 20 10 80 2 38.62 34.67 8135 20 10 90 1.5 52.80 47.14 8138 20 10 90 2 48.80 38.57 814 20 10 80 0.5 50.00 26.00 8141 40 5 80 1.5 38.40 49.65 8144 40 5 80 2 43.64 41.67 8147 40 5 90 1.5 54.07 50.67 8150 40 5 90 2 81.11 80.00 8153 40 10 80 1.5 45.00 54.67 8156 40 10 90 1.5 42.22 62.67 8159 40 10 90 2 41.54 54.67 8162 40 10 80 2 42.22 48.00 8165 60 5 80 1.5 29.60 24.29 8168 60 5 80 2 42.11 41.82 817 20 10 80 0.75 66.67 40.00 8177 60 5 90 1.5 70.59 73.33 * Treatment Moisture ((Kg water/Kg dry straw) x 100) ** Treatment Ammonia Ratio (lb ammonia/lb dry straw) 129 Table 8.6 (Cont’d) Initial conditions Treatment Treatment Treatment Treatment Yield of Yield of ID Moisture time (min) Temperature Ammonia dextrose xylose (%)* (°C) ratio (%) (%) 8180 60 5 90 2 55.38 31.43 8186 60 10 90 2 48.33 57.14 820 20 10 90 0.5 38.18 18.33 824 20 10 90 0.75 60.00 47.50 828 20 10 90 1 40.87 43.33 832 20 5 90 0.1 48.18 28.33 835 20 5 90 0.75 29.09 18.33 841 20 10 80 1 33.57 35.38 845 40 5 80 0.75 82.86 40.00 851 40 5 90 0.5 60.00 27.50 854 40 5 90 0.75 39.13 20.00 857 40 10 80 0.5 46.96 29.23 860 40 10 80 0.75 42.73 30.00 863 40 5 90 1 33.91 36.67 866 40 10 80 1 42.61 48.33 869 40 10 90 0.5 36.00 17.33 872 40 10 90 0.75 40.00 28.33 875 40 10 90 0.1 58.67 65.71 878 60 5 80 0.5 72.31 34.29 88 20 5 ' 80 0.75 40.87 21.54 884 60 5 80 1 51.43 28.33 887 60 5 90 0.5 22 16 890 60 5 90 0.75 33.85 17.33 C172 60 10 80 1.5 46.00 74.00 0184 60 10 90 1.5 32.00 58.46 C49 40 5 80 1 43.64 26.67 untreated rice straw 22.00 12.00 * Treatment Moisture ((Kg water/Kg dry straw) x 100) ** Treatment Ammonia Ratio (lb ammonia/lb dry straw) 130 Table 8.7 MatLab Code for Principal Component Analysis %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% ' %% Matlab code to perform PCA analysis on the data %% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% clear all %% List of files to be read a=dir(’dex*’); b=dir('xyl*'); c=dir(’tot*’); d=[a;b;c]; len=length(d); for j=1zlen fname=d(j).name; %% Read the id and sugar values [id sug] = textread(fname,’%s %f’); f1uor_fname = strcatlid,'.txt’); numbsug = 1ength(sug); index = 0: %% Keep track of ids that do not have a corresponding file miss=fname; for i=1:num_sug f = char(fluor_fname(i)); if (exist(f,'file') ~= 0) x_va1 = textread(f,'%*f %f’); index = index+l; , %% Update x_mat and ygvec with new values x;mat(:,index) = xgval; y_vec(index) = sug(i); else miss = sprintf('%s\n%s',miss,f); end end y;vec = y_vec’; pr=sprintf(’%s: %d', fname,index); displpr) %% Perform PCA [pc score latent tsquare] = princomp(x_mat); %% Only the first 9 components new;pc = pc(1:9.:)'; %% Save the PC oname=strcat(’pc_',fname); save(oname,'new;pc','-double','-ascii',’-tabs') %% Missing ids fp=fopen('missing_files',’a'); 131 fprintf(fp,’%s\n’,miss); fclose(fp); clear(’miss','y_vec',’x_mat’); end LL. 132 BIBLIOGRAPHY 133 9 BIBLIOGRAPHY Agricultural Statistics. 1988. US. Department of Agriculture, Washington, DC. Abraham M., Kurup G. M. (1997). Pretreatment studies of cellulose wastes for optimization of cellulase enzyme activity. Applied Biochemistry and Biotechnology, 62 (2-3), 201-211. Anderson T. H., Weaver F., William F. Owen N. L (1991). Anomalies in diffuse reflectance infrared spectra of wood and wood polymers. J. Mol. Structure, 249, 257-275. Andren R. K., Mandels M. H., Medeiros J. E. (1975). Production of sugars from waste cellulose by enzymatic hydrolysis. App. Polym. Symposium, 28, 205-219. Bacic A., Harris P. J ., Stone B. A. (1988). Structure and function of plant cell walls. In: J. Preiss (Ed) Biochemistry of Plants. Vol 14. Carbohydrates P 529. Academic Press, San Diego, CA. Backa S., Brolin A. (1991). Determination of Pulp Characteristics by Diffuse Reflectance FTIR. Tappi J, 74, 218. Baker A. J. (1973). Effect of lignin on the in vitro digestibility of wood pulp. J. Anim. Sci, 36, 768-771. Beardsley, D. W. (1964). Symposium on Forage utilization: Nutritive value of forage as affected by physical form. Part 11. Beef cattle and sheep studies. Journal of Animal Science, 23, 239-245. Beever, D. E., Osbum D. F., Cammell S. 8., Terry R. A. (1981). The effect of grinding and pelleting on the digestion of italian ryegrass and timothy by sheep. Br. J. Nutr., 46, 357-370. Ben-Ghedalia D., Miron J ., Est Y., Yosef E. (1988). SO; treatment for converting straw into a concentrate-like feed: A growth study with lambs. Anim. Feed Sci. Technol, 19, 401-406. Berber S. A., Rademacher J. P., Sell L. O., Easty DB. (1987). Estimation of lignin in wood pulp by diffuse reflectance Foufier—transfrom infrared spectroscopy. Tappi, 67, 129-133. Berger L. L., Fahey Jr. G. C., Bourquin L. D., Titgemeyer E. C. (1994). Modification of Forage Quality after Harvest, Forage Quality, Evaluation, and Utilization, American Society of Agronomy, Inc., Madision, Wisconsin. 134 Bertran M. 8., Dale B. E. (1985). Enzymatic Hydrolysis and Recrystallization Behavior of Initially Amorphous Cellulose. Biotechnol. Bioeng., 27, 177-181. Beyer M., Steger D., Volker E., Koch H., Fischer K. (1995). Extended Abstracts, Workshop on Li gnocellulosics and Pulp(EWLP). Grenoble, France, ATIP 187. Billa E., Koutsoula E., Koukios E. G. (1999). Flourescence Analysis of Paper Pulps. Bioresource Technol, 67, 25-33. Browning B. L. (1976). Methods of Wood Chemistry. Vol 2, Interscience, New York, NY. Bouveng H. O. (1961). Acta Chem. Scand. 15, 87-96. Carrasco J. E., Saiz M. C., Navarro A., Soriano P., Saez F., Martinez J. M. (1994). Effects of dilute-acid and steam explosion pretreatments on the cellulose structure and kinetics of cellulose fraction hydrolysis by dilute acids in 1i gnocellulosic materials. Appl. Biochem. Biotechnol., 45/46, 23. Castellan A., Davidson R. S. (1994). Steady-State and dynamic fluorescence emission from Abies wood J. Photochem. Photobiol. A: Chem, 78, 275-279. Caulfield D. F., Steffes R. A. (1969). T appi, 52, 1361. Chang H.-M., Allan G. G. (1971). Oxidation. P433485. In K. V. Sarkanen and C. H. Ludwig (ed) Lignins: Occurrence, formation, structure and reactions. Wiley- interscience, New York. Chang V. S., Holtzapple M. T. (2000). Fundamental factors affecting biomass enzymatic I” reacitivity. Appl.Biochem. Biotechnol., 84-6, 5-37. i- Chang V. S., Nagwani M. Holtzapple M. T. (1998). Lime Pretreatment of crop residues bagass and wheat straw. Applied Biochemistry and Biotechnology, 74(3), 135-159. Church D. C. (1991). Livestock Feeds and Feeding, Prentice hall, Englewood Cliffs, New Jersey. Church DC. (1988). The Ruminant Animal Digestive Physiology and Nutrition, Prentice Hall, Inc., New Jersey. Cowling E. B. (1975). Properties of Cellulose and Li gno-cellulosic materials as substrates for enzymatic conversion processes. Biotechnol. Bioeng. Symp., 5,163. 135 Coxworth E., Keman J ., Knipfel J ., Thorlacius O., Crowle L. (1981). Review: Crop residues and forages in western Canada; potential for feed use either with or without chemical or physical processing. Agric. Environm, 6, 245-256. Dale B. E., Leong C. K., Pham T. K., Esquivel V. M., Rios 1., Latimer V. M. (1996). Hydrolysis of Lignocellulosics at low enzyme levels: Application of the AFEX Process. Bioresource Technol., 56, 111-116. Dale 8. E., Moreira M. J. (1983). A freeze explosion technique for increasing cellulose hydrolysis. Biotechnology and Bioengineering Symposium, 12, 31. Dourado F., Gama M. M., Chibowski E., Mota M. (1998). Characterization of cellulose free energy. Journal of Adhesion Science and Technology, 12(10), 1081-1090. Edwards P. B., Mackney A. W. (1939). J. Council Sci. Ind. Res.(Australia), 11, 185-200. EK R., Alderbom G., Nystrom C. (1994). Particle Analysis of Microcrystalline Cellulose-differentiation between individual particles and their agglomerates. International Journal of Pharmaceutics, 111(1), 43-50. EK R., Lennholm H, Davidson R., Nystrom C., Ragnarsson G. (1994). Pore swelling in beads made of cellulose fibers and fiber fragments. International Journal of Pharmaceutics, 111(1), 43-50. Faix O., Bottcher J. O. (1992). The Influence of Particle-size and Concentration in Transmission and Diffuse Reflectance spectroscopy of Wood. Holz als Roh-und Werkstofl; 50, 221. ‘ Fan L. T., Lee Y.-H., Beardmore D. H. (1980). Mechanism of the enzymatic hydrolysis of cellulose: Effects of major structural features of cellulose on enzymatic hydrolysis. Biotechnol. Bioeng., 22, 177-199. Fan L. T., Lee Y.-H., Beardmore D. H. (1981). The influence of major structural features of cellulose on rate of enzymatic hydrolysis. Biotechnol. Bioeng., 23,419. Fan L. T., Gharpuray M. M., Lee Y.-H. (1981). Evaluation of pretreatrnents for enzymatic conversion of agricultural residues. Biotechnol. Bioeng. Symp., 11, 29-45. Feist W. C., Baker A. J ., Tarkow H. (1970). Modifying wood to increase its in vitro digestibility J. Anim. Sci., 30, 832-835. Ferraz A., Baeza J., Rodriguez J., Freer J. (2000). Estimating the chemical composition of biodegraded pine and eucalyptus wood by DRIFT spectroscopy and multivariate analysis. Bioresource Technol., 74, 201-212. 136 Frey-Wyssling A., Muhlethaler K. (1963). Die elementar-fibrillen der cellulose. Makromol. Chem, 62:25. Gharpuray M. M., Lee Y., Fan L. T. (1983). Structural modification of lignocellulosics by pretreatrnents to enhance enzymatic hydrolysis. Biotechnol. Bioeng., 25, 157-172. Ghose, T. K. 1987. Measurement of Cellulase Activities, Pure & Appl. Chem, 59, 257- 268. Grethlein H. E. (1985). The Effect of Pore-size distribution on the rate of enzymatic- hydrolysis of cellulosic substrate. Bio/Technol. 3, 155-160. Grethlein H. E., Converse A. O. (1991). Common aspects of acid prehydrolysis and steam explosion for preheating wood. Bioresourve Technology, 36,77-82. Harrington K. J ., Higgins H. G., Michell A. J. (1964). Infrared spectra of eucalyptus regnaus F. Muell and pinius radiata D. Don. Holzforschung, 18, 108. Hartley W. N. (1893). Observations on the Origin of Colour and of Fluorescence. J. Chem. Soc., 63, 243. Hespell R. B., O’Bryan P. J ., Moniruzzaman M., Bothast R. J. (1997). Hydrolysis by Commercial Enzyme Mixture of AFEX-treated Corn Fiber and Isolated Xylans. Appl.Biochem. Biotechnol., 62, 87-97. Holtzapple M. T., Humphrey A. E. (1984). The effect of Organosolv pretreatment on the enzymatic hydrolysis of poplar. Biotechnology and Bioengineering, 26 (7), 670-676. Holtzapple M. T., Lundeen J. E., Sturgis R., Lewis J ., Dale 8. E. (1992). Pretreatment of lignocellulosic municipal solid waste by Ammonia Fiber Explosion (AFEX). Applied Biochemistry and Biotechnology, 34135, 59-74 Holtzapple M.,T., Jun Jae-Hoon, Ashok G., Patibandala S. L., Dale B. E. (1991). The Ammonia Freeze Explosion (AFEX) Process. Applied Biochemistry and Biotechnology, 28/29, 59-74. Howsman J. A., Marchessault R. H. (1959). The Ball-Milling of cellulose fibers and recrystallization effects. J. Appl. Polym. Sci, 1, 313. Huber T. J. (1981). Upgrading Residues and By-Products for Animals, CRC Press, Inc., Boca Raton, Florida. Iiyama K., Lam B. T., Stone B. A. (1990). Phenolic acid bridges between polysaccharides and 1i gnin in wheat internodes. Phytochem., 29, 733-737. 137 Jung H. G., Allen M. S. (1995). Characteristics of Plant Cell Wall affecting Intake and Digestibility of Forages by Ruminants. J. Anim. Sci, 73, 2774-27 90. Jung, H.G., and Kenneth P. Vogel. (1992). Lignification of switchgrass (Panicum virgatum) and big bluestem (Andropogon gerardii) plant parts during maturation and its effect on fibre degradability. J Sci Food Agric 59:169-176. Jung H. J. 0., Ralph J ., Hatfield R. D. (1991). Degradability of phenolic acid hemicellulose esters: A model system. J. Sci. Food Agric., 56, 469-478. Kansas City Weekly F eedstufi Review, Jun 22, 1999. Kerley M. S., Fahey G. C. Jr., Gould J. M., Iannotti E. L. (1988). Effects of lignificatin, cellulose crystallinity and enzyme accessible space on the digestibility of plant cell wall carbohydrates by the ruminant. Food Microstructure, 7, 59—65. Knappert D., Grethlein H., Converse A. (1980). Partial Acid Hydrolysis of Cellulosic Materials as a Pretreatment for Enzymatic Hydrolysis. Biotechnology and Bioengineering, 22, 1449-1463. Kogel-Knabnerl (2000). Analytical approaches for characterizing soil organic matter. Organic Chemistry, 31(7-8), 609-625. Kong R., Engler C. R., Soltes E. J. (1992). Effect of Cell-Wall Acetate, Xylan backbone and Li gnin on enzymatic hydrolysis of Aspen Wood. Appl.Biochem. Biotechnol., 34, 23- 35. Koullas D. P., Christakopuolos P., Kekos D., Macris B. J ., Koukios E. G. (1992). Correlating the Effect of Pretreatment on the Enzymatic Hydrolysis of Straw. Biotechnol. Bioeng. 39, 1 13-116. Lam T.B.T., Iiyama K. and Stone B.A. (1990). Distribution of free and combined phenolic acids in wheat intemodes. Phytochemistry 29: 429-433. Lamola A. A., Hammond G. S., Mallory F. B. (1965). The Blue Emission of cis- stilbenes. Photochem. Photobiol., 4, 259. Lindgren T., Edlund U. (1998). Prediction of lignin content and pulp yield-from black liquor composition using near-infrared spectroscopy and partial least squares regression. Nordic Pulp and paper Research Journal, 13(1), 76-80. Love G. D., Snape C. E., Jarvis M. C. (1992). Determination of the Aromatic Content in Oak Wood by quantitative solid-state c-l3-NMR. Biopolymers, 32(9), 1187-1192. Males, J. R. (1987). Optimizing the utilization of cereal crop residues for beef cattle. J. Anim. Sci, 65, 1124-1130. 138 Mandels M., Hontz L., Nystrom J. (1974). Enzymatic hydrolysis of Waste Cellulose. Biotechnol. Bioeng., 16, 1471. Mandels M., Stemberg D. (1976). Recent Advances in Cellulase Technology. J. Ferment. Technol., 54, 4; 267. Mason V. C., Hartley R. D., Keene A. S., Cobby J. M. (1988). The effect of ammoniation on the nutritive value of wheat, barley and cat straws. 1. Changes in chemical composition in relation to digestibility in vitro and cell wall degradability. Anim. Feed. Sci. Technol., 19, 159-171. Mes-Hartree M., Dale B. E., Craig W. K. (1988). Appl. Microbiol Biotechnol. 29, 462. Michell A. J. (1994). Infra-red spectroscopy transformed-new applications in wood and pulping chemistry. Appita 47, 29-37. Miller, G. L. 1959. Use of Dinitrosalicylic Acid Reagent for Determination of Reducing Sugar. Anal. Chem. 31, 426-428. Minson D. J. (1990). Forage in Ruminant Nutrition, Academic Press, Inc., New York. Minson, D. J. (1963). The effect of pelleting and wafering on the feeding value of roughage-A review. J. Br. Grassl. Soc. 18, 39-44. Moore, L. A. (1964). Symposium on forage utilization: Nutritive value of forage as affected by physical form. Part I. General principles involved with ruminants and effect of feeding pelleted or wafered forage to dairy cattle. Journal of Animal Science, 23, 230- 238. Morris P. J ., Mowat D. N. (1980). Nutritive value of ground and /or ammoniated corn stover. Can. J. Anim. Sci, 60,327-336. Muller M., Czihak C., Schober H., Nishiyama Y., Vogl G. (2000). All disordered regions of native cellulose show common-low frequency dynamics. Macromolecules, 33(5), 2000. Nevell, T. P. (1985). Oxidation of cellulose. P. 243-265. In T. P. Nevell and S. H. Zeronian (ed) Cellulose chemistry and its applications. Ellis Horwood Ltd., Chichester. Nesse N., Wallick J., Harper J .M. (1977). Pretreatment of Cellulosic Wastes to Increase Enzyme reactivity. Biotechnol. Bioeng., 19, 323. Norkrans B. (1950). Studies in growth and cellulolytic enzymes of Tricholoma. Physiol. Plant, 3, 75. 139 Ogier J. C., Ballerini D., Leygue J. P., Riga] L., Pourquie J. (1999). Ethanol Production form lignocellulosic biomass. Oil and Gas Science and Technology, 54(1), 67-94. Olmstead J. A., Gray D. G., (1993). Fluorescence Emission from mechanical pulp sheets. J. Photoch. Photobio. A., 73(1), 59-65. Olmstead J. A., Gray D. G., (1997). Fluorescence Spectroscopy of Cellulose, Lignin and Mechanical Pulps: A Review. Journal of Pulp and Paper Science, 23, J57l-J579. Owen N. L., Thomas D. W. (1989). Infrared studies of hard and soft woods. Appl. Spectroscopy 43, 451-455. Pandey K. K. (1999). A study of chemical structure of soft and hardwood and wood polymers by FTIR spectroscopy. J. Appl. Polym. Sci, 71, 1969-1975. Pandey K. K., Theagarajan K. S. (1997). Holz als Roh-und Werkstofl', 55, 383. Paralikar K. M., Betrabet S. M. (1977). Electron diffraction technique for determination of cellulose crystallinity. Journal of Applied Polymer Science, 21(4), 899-903. Puri V. P. (1984). Effect of cystallinity and degree of polymerization of cellulose on enzymatic sacchrification. Biotechnol. Bioeng., 26, 1219-1222. Ralph J ., Hatfield R. D., Quideau S., Helm R., Grabber J. H, Hans-Joachim, Jung 6., J. (1994). Pathway of p-coumaric acid incorporation into maize lignin as revealed by NMR. American Chem. Soc., 116, 9448-9456. Rivers D. B., Emert G. H. (1988). Factors affecting the enzymatic hydrolysis of bagasse and rice straw. Biotechnol. Bioeng., 31, 278-281. Rodrigues J ., Faix O., Pereira H. (1998). Determination of lignin Content of Eucalyptus Globulus Wood using FTIR spectroscopy. Holzforschung, 52 (1), 46-50. Rowland S. P. (1975). Selected aspects of structure and accessibility of cellulose as they relate to hydrolysis. Biotechnol. Bioeng. Symp., 5, 183-191. Sasaki T., Tanaka T., Nanbu N., Sato Y., Kainuma K. (1979). Correlation between X-ray diffraction measurements of cellulose crystalline structure and the susceptibility to microbial cellulase. Biotechnol. Bioeng., 21, 1031-1042. Sega] L., Creely J. J ., Martin Jr. A. E., Conrad C. M. (1959). An emperical method for estimating the degree of crystallinity of native cellulose using the x-ray diffractor. Textile Res. J. 29, 786-794. 140 Silva J.C.E., Nielson B. H., Rodrigues J ., Pereira H., Wellendorf H. (1999). Rapid determination of 1i gnin content in sitka spruce wood by fourier transform infrared spectrometry. Holzforschung, 53(6), 597-602. Stewart D., Morrison 1. M. (1995). Determination of lignin in the prescence of ester- bound substitued cinnamic acids by a modified acetyl bromide procedure. J. Sci. Food Agric., 69, 151-157. Stewart D., Wilson M.H., Hendra P. J ., Morrison I. M. (1995). Fourier-transform infrared and raman spectroscopic study of biochemical and chemical treatments of oak wood and barley straw. J. Agric. Food Chem, 43, 2219-2225. Stone J. E., Scallan A. M., Donefer E., Ahlgrer E. (1969). Cellulase and Their Applications. Gould R. G., ed., American Chemical Society, Washington D. C., p. 219- 241. Sullivan J. T. (1959). A rapid method for the determination of acid-insoluble lignin in forages and its relation to digestibility. J. Anim. Sci, 18, 1292-1298. Sundstol F. (1988). Straw and other Fibrous by-products. Livest. Prod. Sci, 19 (1-2), 137-158. Sundstol F., Coxworth E. M. (1984). Ammonia treatment. p. 196-247. In F. Sundstol and E. Owen (ed.) Straw and other fibrous by-products as feed. Elsevier Science Publishers, Amsterdam, The Netherlands. Thompson D. N., Chen H.-C., Grethlein H. E. (1992). Comparision of Pretreatment Methods on the basis of Available Surface area. Biores. T echnol., 39,155-163. Tsao G. T., Ladisch M., Ladisch C., Hsu A. T., Dale B., Chou T. (1978), Annual Reports on Fermentation Processes, vol. 2, D. Perhnan, Ed. (Academic, New York,), p.1. T eixeira L. C., Linden J. C., Schroeder H. A. (1999). Optimizing peracetic pretreatment conditions for improved simultaneous saccharaification and co—fermentation (SSCF) of sugar cane bagasse to ethanol fuel. Renewable energy, 16(1-4), 1070-1073. Tylli H., Forsskahl 1., Olkkonenc (1995). The effect of Photoirradiation on High Yield Pulps: Spectroscopy and Kinetics. J. Photochem. Photobiol. A. Chem, 87, 181. Tylli H., Forsskahl I., Olkkonenc (1996). Photochromic Behavior of Ozonated and Photoirradiated Cellulose Studied by Fluorescence Spectroscopy. Cellulose, 3, 203. Van Soest P. J. (1969). Cellulase and Their Applications. Gould R. G., ed., American Chemical Society, Washington D. C., p. 262-278. 141 Van Soest P. J. (1982). Nutritional Ecology of the Ruminant. p. 125-126. O&B Books, Inc., Corvallis, OR. Van Soest P. J ., Wine R. H. (1968). Determination of lignin and cellulose in acid- detergent fiber with permanganate. J. Assoc. Oflicial Anal. Chem, 51, 780. Vincent J. H. Sewalt, Fontenot P. J. Allen V. G., Glasser W. G. (1996). Fiber composition and in vitro digestibility of corn stover fractions in response to ammonia treatment. American Chemical Society. J. Agric. Food Chem, 44(10). Vinzant T. B., Ehrrnan C. I., Adney Wa. S., Thomas S. R., Himmel M. E. (1997). Simultaneous saccharification and fermentation of pretreated hardwoods. Appl. Biochem. Biotechnol., 62, 99-104. Wang L., Dale B. E., Yurttas L., Goldwasser I. (1998). Cost Estimates and Sensitivity Analyses for the Ammonia Fiber Explosion Process. Appl. Biochem. Biotechnol., 70-72, 5 1-66. Wei C.-J., Cheng C.-Y. (1985). Effect of hydrogen peroxide pretreatment on the structural features and the enzymatic hydrolysis of rice straw. Biotechnol. Bioeng., 27, 1418-1426. Weimer P. J ., Weston W. M. (1985). Relationship between the fine structure of native cellulose and cellulose degradability by the cellulase complexes of Trichoderma reesei and clostridium thermocellum. Biotechnol. Bioeng., 27, 1540-1547. Yu X. C., Atalla R. H. (1998). A Staining technique for evaluating the pore structure variation of microcrystalline cellulose powders. Powder Technology, 98(2), 135-138. 142