Dilute sulfuric acid pretreatment of switchgrass in microwave reactor for biofuel conversion: An investigation of yields, kinetics, and enzymatic digestibility of solids
v TABLE OF CONTENTS
Page List of Tables Vii List of Figures X List of Abbreviations Xi Abstract Xii Background & Significance 1 Lignocellulosic biomass 4 Cellulose 6 Hemicellulose 7 Lignin 8 Switchgrass 9 Pretreatment 11 Chemical pretreatment 13 Acid hydrolysis 15 Conventional heating 18 Mechanism 19 Switchgrass pretreated using conventional heating 19 Microwave heating 20 Mechanism 21 Switchgrass pretreated using microwave heating 24 Materials & Methods 26 Materials 26 Substrates 26 Acid 27 Cellulase enzyme 27 Pretreatment 27 Conventional heated reactor 28 Microwave heated reactor 29 Enzymatic hydrolysis 30 Analysis 30 Experimental Design 32 Results and Discussion 35 Pressure 35 Biomass 36 Mass loss 36 Color 40 Porosity 41 Pretreatment liquor 43 PH 43 Glucose 46 Xylose 49 Degradation products 50 Enzymatic hydrolysis liquor 55 Glucose yield as a function of pretreatment conditions 55
Glucose yield as a function of pretreated biomass composition 57 Model 60 Combined severity factor 60 Glucose yield 60 Xylose yield 62 Degradation product yield 62 Kinetic model 63 Glucose yield in the pretreatment liquor 66 Xylose yield in the pretreatment liquor 67 Degradation product yield in the pretreatment liquor 68
Overall Mass, Energy, & Economic Analysis 70 Mass balance 70 Energy balance 71 Economic analysis 73 Harvest 74 Delivery 75 Milling 76 Pretreatment Continuous/Batch, Conventional/Microwave 77 78 Enzymatic hydrolysis 82 Waste stream outlet 84 Financial summary- Paypack period / Net present value 86 Outlook 90 Conclusion 91 References 96 Appendix 1- Figures 102 Appendix 2- Methods 133 Determination of carbohydrates in biomass by high performance liquid chromatography 133 Determination of structural carbohydrates and lignin in biomass 144 Determination of sugars, byproducts, and degradation products in liquid fraction process samples 152 Enzymatic saccharification of lignocellulosic biomass 164 Pretreatment reactor protocol 170 Vita 173
vii List of Tables
Page 1 Switchgrass forage yield cited in the literature 11 2 Conventional pretreated switchgrass in the literature 20 3 Microwave pretreated switchgrass in the literature 25 4 Pretreatment experimental design 33 5 Final reactor pressure obtained during experimentation 36 6 Mass loss result summary 39 7 Pretreated biomass composition result summary 39 8 pH result summary 45 9 Change in pretreatment liquor pH as a function of pretreatment parameters 46 10 Pretreatment liquor glucose result summary 48 11 Pretreatment liquor xylose result summary 48 12 Pretreatment liquor hydroxy-methyl furfural (HMF) result summary 54 13 Pretreatment liquor acetic acid result summary 54 14 Enzymatic hydrolysis liquor glucose result summary 59 15 Glucose in switchgrass pretreatment-liquor as a function of combined severity factor
61 16 Glucose from switchgrass enzymatic hydrolysis as a function of combined severity factor 61 17 Kinetic constants for the glucose formation in the switchgrass pretreatment liquor 66 18 Kinetic constants for the xylose formation in the switchgrass pretreatment liquor 67 19 Kinetic constants for the HMF formation in the switchgrass pretreatment liquor 68 20 Kinetic constants for the acetic acid formation in the switchgrass pretreatment liquor 69 21 Mass balance for the pretreatment process 71 22 Energy balance for the pretreatment process 72 23 Seed price for selected perennial grasses 74 24 Total feedstock cost 76 25 Pretreatment chemical cost 77 26 Investment and operating cost for conventional batch pretreatment 80 27 Investment and operating cost for conventional continuous pretreatment 80 28 Investment and operating cost for microwave batch pretreatment 82 29 Investment and operating cost for microwave continuous pretreatment 83 30 Major operating conditions for enzymatic hydrolysis 83
31 Investment and operating cost for enzymatic hydrolysis 84 32 Waste stream potential 86 33 Financial summary for the pretreatment reactor systems 88 34 Payback period analysis 89 35 Net present value analysis 90
List of Figures
Page 1 Lignocellulosic structure 6 2 Cellulose structure 7 3 Hemicellulose structure 8 4 Lignin structure 9 5 Switchgrass 10 6 Cellulose to glucose reaction 17 7 Microwave mechanism 22 8 Microwave vs. conventional heating 24 9 Experimental switchgrass 27 10 PARR® High Temperature High Pressure Model 4575A 28 11 CEM Explorer 48 Microwave Reactor 29 12 Process flow diagram 34 13 Switchgrass discoloration due to pretreatment 41 14 SEM photograph of unpretreated switchgrass 42 15 SEM photograph of conventional-pretreated switchgrass 42 16 SEM photograph of microwave-pretreated switchgrass 43
A1 PARR® reactor pressure as a function of temperature and ramp time 102 A2 CEM Explorer reactor pressure as a function of temperature and ramp time 102 A3 Avicel® mass loss as a function of conventional and microwave pretreatment conditions 103 A4 Whatman paper mass loss as a function of conventional and microwave pretreatment conditions 104 A5 Switchgrass mass loss as a function of conventional and microwave pretreatment conditions 105 A6 Switchgrass cellulose as a function of conventional and microwave pretreatment conditions 106 A7 Switchgrass xylan as a function of conventional and microwave pretreatment conditions 107 A8 Avicel® liquor pH as a function of conventional and microwave pretreatment conditions 108 A9 Whatman paper liquor pH as a function of conventional and microwave pretreatment conditions 109 A10 Switchgrass liquor pH as a function of conventional and microwave pretreatment conditions 110 A11 Glucose in Avicel® liquor as a function of conventional and microwave pretreatment conditions 111 A12 Glucose in Whatman paper liquor as a function of conventional and microwave pretreatment conditions 112 A13 Glucose in switchgrass liquor as a function of conventional and 113
x microwave pretreatment conditions A13A Glucose in switchgrass liquor as a function of conventional and microwave combination pretreatment conditions 114 A14 Xylose in switchgrass liquor as a function of conventional and microwave pretreatment conditions 115 A15 HMF in Avicel® liquor as a function of conventional and microwave pretreatment conditions 116 A16 HMF in Whatman paper liquor as a function of conventional and microwave pretreatment conditions 117 A17 HMF in switchgrass liquor as a function of conventional and microwave pretreatment conditions 118 A17A HMF in switchgrass liquor as a function of conventional and microwave combination pretreatment conditions 119 A18 Acetic acid in switchgrass liquor as a function of conventional and microwave pretreatment conditions 120 A19 Xylitol in switchgrass liquor as a function of conventional and microwave pretreatment conditions 121 A20 Succinic acid in switchgrass liquor as a function of conventional and microwave pretreatment conditions 122 A21 Glucose in enzymatic hydrolysis liquor as a function of conventional and microwave pretreatment conditions 123 A21A Glucose in enzymatic hydrolysis liquor as a function of conventional and microwave combination pretreatment conditions 124 A21B Normalized glucose yield as a function of conventional and microwave combination pretreatment conditions 125 A22 Glucose as a function of pretreated biomass cellulose content for conventional and microwave reactors 126 A23 Glucose in enzymatic hydrolysis liquor as a function of pretreated biomass cellulose and lignin content for conventional and microwave reactors 127 A24 Glucose in the switchgrass-pretreatment liquor as a function of combined severity factor (CSF) 128 A25 Glucose in switchgrass-enzymatic hydrolysis liquor as a function of combined severity factor (CSF) 128 A26 Combined glucose (pretreatment and enzymatic hydrolysis liquors) as a function of combined severity factor (CSF) for conventional and microwave reactors 129 A27 Xylose in switchgrass pretreatment liquor as a function of combined severity factor (CSF) for conventional and microwave reactors 130 A28 HMF in all pretreatment liquors as a function of combined severity factor (CSF) for conventional and microwave reactors 130 A29 Acetic acid in switchgrass pretreatment liquors as a function of combined severity factor (CSF) for conventional and microwave reactors 131 A30 Mass and energy balance 132
List of Abbreviations
BTU British thermal units CSF Combined severity factor HMF Hydroxy-methyl-furfural HPLC High Performance Liquid Chromatography LAP Laboratory Analytical Procedure PSI Pounds per square inch SEM Scanning electron microscope NPV Net present value
Lignocellulosic materials provide a raw material source for biofuel conversion and offer several advantages over fossil fuels- usage of a renewable resource, reduced greenhouse emissions, a decreased dependence on foreign oil, and stimulation of the agricultural sector. However, a primary technological challenge in converting lignocellulosic biomass into fuel is overcoming the recalcitrance of its matrix to enzymatic hydrolysis. To overcome these problems for chemical processing, naturally occurring cellulose biomass must be pretreated before it can be further processed using enzymatic hydrolysis or bioconversion.
The goal of this work was to develop a model that predicts the glucose yield (pretreatment and enzymatic digestibility) of dilute acid pretreated switchgrass as a function of pretreatment process conditions (acid loading, 0-1.5 vol%, temperature, 165- 195 o C, and residence time, 1-10 min). This project was the first study that used a multi- variable design experimental series to directly compare the pretreatment effectiveness (product yield, biomass composition and appearance, pH, etc) of using conventional and microwave heated reactors.
Microwave-pretreated switchgrass afforded up to a 100% higher total glucose yield (combined pretreatment and enzymatic-hydrolysis liquor yields) at equivalent pretreatment severity and at one tenth of the reaction time, relative to conventional
pretreatment. Under best pretreatment conditions of 0.75 vol% acid, 195 o C, 1 min residence time, 99% glucose yield and 99% hemicellulose removal were achieved.
Kinetic parameters were estimated for the cellulose and xylan hydrolysis reactions in the pretreatment liquor and the solid residue. The kinetic model gave an average correlation coefficient of 0.93 for all reactions. In addition, the combined severity factors (CSF) were also determined for each experiment. Highest observed enzymatic glucose yield corresponded to a CSF of 1.7.
A mass and energy balance, and economic analysis based on production scale was developed for both reactor systems. The microwave pretreatment process theoretically yielded 48% more ethanol relative to the conventional process. For microwave pretreatment to be commercially viable, two criteria must be met. One, the cost for large- scale continuous microwave reactors would need to be significantly lower than current estimates. And second, higher solids content must be used (> 20 wt% in the slurry) to maximize output.
1 1.0 BACKGROUND AND SIGNIFICANCE
Gasoline is a petroleum-derived liquid mixture consisting of 5-to-12-carbon hydrocarbons, including parrafins, naphthenes, aromatics, olefins, and hazardous chemicals (5 to 35 percent by volume) such as benzene (to increase the octane rating), toluene, naphthalene, trimethylbenzene, and methyl tert-butyl ether (MTBE) (Kaufmann and Shiers, 2008).
Global petroleum consumption has reached 84,035,000 barrels per day, with U.S. petroleum consumption at 20,802,000 barrels per day. Current U.S. motor gasoline consumption is 384.7 million gallons per day, or 140 billion gallons annually. The US is set to consume 290 billion gallons of gasoline a year in cars and trucks by 2050. Inflation adjusted gasoline prices have skyrocketed from $1.35 to $3.22 per gallon from 1998 to 2008. (Energy Information Administration, 2008)
Worldwide energy consumption for 2007 was approximately 5x10 17 BTUs (British Thermal Units) according to the Energy Information Administration (Energy Information Administration). The US accounts for about 27% of this consumption (Energy Information Administration, 2008a). The agency projects global energy consumption to surpass 7x10 17 BTUs by 2030. More than 50% of the projected increase in global energy demand over the next twenty years is attributed to the growing economies of China and India, which currently account for approximately 18% of global
2 energy consumption. This increase offsets the 17% projected decline in the US share of global energy consumption by 2030. (Energy Information Administration, 2008)
In the long term, fossil fuels are not projected to satisfy the growing global energy demand. Many industry experts predict that the world will face a “peak oil” situation within the current century. Estimates on the data for “peak oil” vary from 2010 to 2030. Models by Campbell and Laherrere (1998), USGS (2000), IEA World Energy Outlook (Energy Information Administration, 2008), and Jackson (2007) alternatively project peak oil to arrive by 2010, 2023, 2030, and after 2030, respectively. Differences in the estimated dates for peak oil result from varying estimates of the magnitude of untapped reserves. Current estimates for crude oil long-term availability range from 0.8 to 2.9 x10 12 barrels (Kaufmann and Shiers, 2008).
There is tremendous interest in the commercialization of alternatives to petroleum-derived fuels. This is a direct result of the increasing global energy demand, uncertainty of crude oil supplies, and environmental impacts from the use of these fossil fuels. In addition, there is also concern about US dependence on the use of foreign oil supplies and the price fluctuations caused by geo-political situations. One example is the 1973 Arab oil embargo, which resulted in spikes in crude oil prices four times over a 12- month period. This resulted in a US recession, and a 3% decline in the US gross domestic product (Hirsch, 2008).
3 Studies have shown that global climate change is a result of forced warming due to greenhouse-gas emissions (Hegerl et al., 2007). These greenhouse gases (i.e., carbon dioxide, methane, and nitrous oxide), account for more than 50% of the overall greenhouse effect and are liberated by fossil fuel combustion (Schnoor, 2005). Therefore, the projected increase in energy demand will result in an increased use of fossil fuels and greenhouse emissions. Carbon dioxide emissions are projected to increase from 2x10 10
tons in 1990 to over 4x10 10 tons by 2030 (Energy Information Administration, 2008a). Sulfur and nitrous oxide emissions are other byproducts of fossil fuel combustion. These gases are major contributors to acid rain, which is harmful to freshwater sources, forests, soils, and buildings, in addition to adversely affecting human health (Demirbas, 2004).
Coal and crude oil together represented over 60 percent of domestic energy consumption in 2007. Approximately 60% of the total crude oil in the US is refined into motor gasoline. Renewable energy represents less than eight percent, with only half obtained from biomass. However, 9.2 percent of energy usage in Europe is derived from renewable resources, with some countries using as much as 41 percent. (Energy.eu, 2006). The Department of Energy (DOE) and the US Department of Agriculture (USDA) have both reported that over 1x10 19 tons of biomass can be harvested to displace up to 30 percent of current fossil fuel usage (Perlack et al., 2005).
A comprehensive renewable energy plan is necessary to the meet the projected global energy usage and address environmental concerns associated with fossil fuels.
4 Renewable energy sources such as biomass, geothermal, hydroelectric, solar, and wind are important parts of an environmentally sustainable energy plan.
Biofuels (e.g., bioethanol, biodiesel, and biobutanol) play a key role in this energy plan. Biofuel are produced by the process of converting organic matter into a combustible fuel as a replacement for fossil fuel. This replaces oil and natural gas, focusing on the use of organic matter in the efficient production of liquid and gaseous biofuels, which yield high net energy gains. This alternative fuel source can be derived from biomass, which is a readily renewable energy source, unlike other natural resources such as petroleum, coal, or nuclear fuels. They offer several advantages over fossil fuels: usage of a renewable resource, reduced greenhouse emissions, decreased dependence on foreign oil, and stimulation of the agricultural sector (Sun, 2005). These alternatives have the potential to replace a significant amount of gasoline in the transportation sector.
1.1 Lignocellulosic biomass
Biomass consists of harvested plant-derived materials that are abundant, inexpensive, and potentially convertible to fuel by fermentation processes. The material can be found as starch in corn, wheat, potatoes, cassava, and other agricultural products and as monomeric sugars or soluble oligomers in corn syrup, molasses, raw sugar juice, sulfite waste liquors. (Ng, 1983)
5 Current energy-crop production competes for fertile land with food (corn, rice, sugar, and wheat) and their residues (e.g., corn stover or soybean hulls). This also increases pollution from fertilizers and pesticides, and is harmful to the biodiversity of the land (Tilman, 2006). One primary objection to food-based energy crop production is that it could divert agricultural production away from food crops. This could lead to greater food shortages in both the poor and developed countries. There was a 20-million- ton increase in world grain consumption in 2007, roughly 1%. A large component of that – 14 million tons – was used to fuel cars in the U.S. This leaves only six million tons to cover growing food needs. (US Department of Agriculture, 2007) The key to lessening demand for grain is to commercialize biofuel production from low-input crops such as lignocellulosic biomass in the form of perennial grasses, wood chips, crop residues, forest and mill residues, and urban refuse. (Ng, 1983).
Naturally occurring lignocellulosic materials, as shown in Figure 1, have carbohydrate-rich cellulose and hemicellulose fibers that are surrounded by a lignin seal. This forms a complex structural matrix that is resistant to enzymatic hydrolysis. The hemicellulose fibers act like a glue that fill the voids between and around cellulose and hemicellulose fibers. The lignin acts as a protective sheath, thus providing the rigid characteristics. This structure reduces accessibility to the polysaccharide molecules. Hence, removal of the hemicellulose and lignin greatly enhances polysaccharide accessibility. The carbohydrate and lignin composition differs based on the plant species (Sun, 2005).
Figure 1: Lignocellulosic structure
In addition to the lignin seal, cellulose chains are held together laterally by intermolecular hydrogen bonds (Fengel and Wegener 1984). These intramolecular hydrogen bonds form between repeating glucose units (Fengel and Wegener 1984). The combined effect of the bonding energies of the hydrogen bonds increases the rigidity of cellulose, causing further insolubility and resistance to hydrolysis.
Cellulose fibers are highly stable homopolymer chains of β-D-glucose units that are linked via β-1-4 glycosidic bonds. The basic repeat unit of cellulose is cellobiose, which consists of two glucose molecules. This linearity of the cellulose chains results in a highly ordered packing of cellulose chains that interact via inter- and intra-molecular
7 hydrogen bonds involving the hydroxyl groups and hydrogen atoms of adjacent glucose units. As a result, cellulose fibers contain both crystalline fibers and some amorphous regions. In a biomass feedstock, cellulose is the primary reservoir of glucose, the desired fermentation substrate. However, overcoming the crystallinity of the cellulose fibers is a major obstacle for efficient enzymatic hydrolysis (Fengel and Wegener 1984).
Figure 2. Cellulose structure
Hemicellulose is an amorphous biopolymer. These heteropolymer fibers vary in structure and composition, and are composed of five-carbon sugars such as xylose and arabinose, and six-carbon sugars such as galactose and mannose. Switchgrass contains two primary types of hemicellulose: arabinoxylan and glucomannan. Arabinoxylan, which consists of a xylan backbone made up of β-1,4-linked D-xylose units with frequent arabinose side chains, is the dominant hemicellulose component (Fengel and Wegener 1984). The presence of arabinose side chains reduces hydrogen bonding, which contributes to the low crystallinity of hemicellulose. Glucomannan is the minor
8 hemicellulose component. This component is a copolymeric chain of glucose and mannose units. Intermittent branching in glucomannan also contributes to the low crystallinity (Fengel and Wegener 1984).
Figure 3. Hemicellulose (xylan) structure
Lignin is a stable, high-molecular-weight compound built of phenylpropane units: p-coumaryl alcohol, coniferyl alcohol, and synapyl alcohol. These units are referred to as monolignols. Lignin has a highly complex structure and is difficult to illustrate as basic structural units. The proportions of these components vary based on the type of lignocellulosic material. Switchgrass is comprised of equal portions of all three monolignols. There are many types of carbon-carbon and ether bonds between individual monolignols. As a result, a complex lignin structure consisting of dimers, trimers, and tetramers is formed by random linkages. The carbon-carbon bonds are the strongest, contributing the major part of the barrier nature of lignin (Fengel and Wegener 1984).
Figure 4. Lignin structure
To be sustainable, biomass production must not interfere with existing food-crop production. One means of addressing this is to grow and harvest biomass must be harvested on marginal lands not currently in production. There are approximately 202 million acres of agriculturally abandoned and degraded land in the U.S. that can be used to grow energy crops such as perennial grasses (Tilman, 2006). These grasses are commonly used as fodder crops, and contribute to the energy supply on farms through the use of draft animals (Lewandowski, 2003). Perennial grass is one energy-crop candidate that can be produced on most agricultural land resources, many of which are not suitable for row crops. These grass crops have the potential to achieve high growth rates on more marginally productive croplands where erosion is a concern and soil stabilization is needed (Tolbert, 1998) This development also has the potential for stimulating the agricultural sector by providing a new source of income for farmers (Alizadeh, 2005).
Switchgrass (Panicum virgatum, L., Poaceae), as shown in Figure 5, is a warm- season, sod-forming, tall grass, which combines good forage attributes and soil- conservation benefits. This North American native perennial grass belongs to the subfamily Panicoideae of the Gramineae family. This species is commonly associated with the natural vegetation of the Great Plains and the western Corn Belt and is widely distributed in grasslands and non-forested areas throughout North America east of the Rocky Mountains. This grass has been planted in pasture and range-grass mixtures for many years and has become increasingly important as a pasture grass because of its ability to be productive during the hot months of summer, when cool-season grasses are less productive. In southern parts of the US, switchgrass can grow to more than three meters tall and develop roots to a depth of more than 3.5 m (Blake, 2008).
Figure 5. Switchgrass Source: (Elberson, 2009)
Switchgrass can be harvested in a variety of soil types. Further, it is heat and drought tolerant, while growing well on soils that are shallow and rocky. It is also tolerant to wet areas, environmental restoration, and crop-management treatments. Switchgrass can be easily integrated into existing farming operations because conventional equipment for seeding, crop management, and harvesting can be used. This grass can grow on sand to clay loam soils and can tolerate soils with pH values ranging from 4.9 to 7.6. Annual yields have been reported to be between 11.1 and 34.6 Mg dry mass per hectare (Lewandowski, 2003). Blake (2008) reported that switchgrass can yield between 500 and 1,000 gallons of ethanol per acre using existing technology.
Table 1. Switchgrass forage yield cited in the literature Reference Region Yield, Mg ha - 1
Lewandowski et al. Texas 13.2 Lewandowski et al. Upper South 12.1 Lewandowski et al. Alabama 26.0-34.6 Lewandowski et al. Britain 11.1
A primary technological challenge in converting lignocellulosic biomass into fuel is overcoming the recalcitrance of its matrix to enzymatic hydrolysis. To overcome these problems for chemical processing, naturally occurring cellulosic biomass must be pretreated before it can be enzymatically hydrolyzed. Pretreatment is one of the most expensive and least technologically mature conversion steps in the cellulosic ethanol process (Laser, 2001). The purpose is to transform the lignocellulosic structure into a usable fermentation substrate. Economic viability of the pretreatment process depends on
its ability to minimize energy demands and limit costs, such as feedstock size reduction, materials of construction, and treatment of process residues (Mosier, 2003).
To qualify as effective, a pretreatment must meet the following criteria: 1) it maximizes the fermentable glucose yield, 2) it minimizes the formation of fermentation inhibitors from sugar degradation, and 3) it is economically efficient. Principal substrate factors that have been correlated with pretreatment effectiveness include increased cellulose pore volume and hemicellulose and lignin removal.
Pretreatment processes can be loosely grouped into three categories: physical, microbial, and chemical. Physical pretreatments, which demand large amounts of energy, employ purely mechanical means to reduce feedstock particle size, thus increasing surface area available for enzymatic hydrolysis. Examples of such processes include ball milling and compression milling. The primary issue associated with physical pretreatments is the relatively high energy cost. Microbial pretreatment uses microorganisms to remove lignin and improve enzymatic cellulose digestibility. An example of such processes is the use of the fungus Cyathus stercoreus to improve hydrolysis. The primary issues associated with microbial pretreatment include slow kinetic and high economic considerations (Hu, 2007). Chemical pretreatments use a variety of chemicals as pretreatment agents: water, acids, alkalis, organic solvents, oxidizing agents, and supercritical fluids. Dilute acid, liquid anhydrous ammonia, lime, and ionic solvent pretreatments have emerged as particularly effective chemical methods (Laser, 2001).
1.3.1 Chemical pretreatment
Chemical pretreatment has been a widely explored approach to overcoming the recalcitrance of natural biomass. Many acids, bases, and other chemicals promote hydrolysis and improve fermentable sugar yield through the removal of hemicellulose and/or lignin. An extensive array of chemical pretreatment options such as the use of oxidizing agents, acids, bases, and other solvents have been investigated. Oxidizing agents tested include alkaline peroxide, sodium and calcium hydroxide, ozone, dioxane, and peroxyacid (organosolv). Acids evaluated include sulfuric acid, hydrochloric acid, phosphoric acid, and nitric acid. Chemical solvents such as ammonia, aprotic solvents (i.e., DMSO), and metal complexes have been explored. These chemicals have shown varying degrees of effectiveness in reducing cellulose crystallinity, disrupting the lignin matrix, and dissolving cellulose (Hu, 2007).
Reaction time, together with temperature and pH, has been reported to influence the pretreatment severity or harshness. Several studies expressed pretreatment severity in terms of a combined severity factor (CSF), that account for multiple process conditions. (Schell, 2003; Kabel, 2006; Chum, 1990) The CSF can be used to determine the best set of experimental parameters required to balance the maximization of hemicellulose and lignin removal with the minimization of glucose degradation, enabling further use of the remaining cellulose (Garrote, 1999). The proposed severity factor is based on an approximation to Arrhenius temperature behavior, but is not limited to first-order kinetics
and allows the well-known reduction in reaction rate with extent of reaction to be accommodated. The formalism presented here linearizes the temperature behavior for convenience, and is equivalent to the Arrhenius formal treatment. The CSF provides a method for consolidating the effects of pretreatment temperature, residence time, and acid concentration into a single parameter, which can be useful for analyzing results. This factor is dependent on process conditions, and does not reflect any physical parameter. CSF is calculated by equation 1:
(1) pHetCSF T − ×= − 75.14 100 10 log
where t is the reaction time in minutes, T is the reaction temperature in degrees Celsius, and pH is the final pH of the pretreatment liquor. This equation is based on several assumptions. First, the practical operating range span –4 to 3, with highest observed hemicellulose removal at CSF values between 1.4 and 1.7 (Schell, 2003). Low calculated CSF values (-4 to 0) represent less harsh conditions (i.e. relatively low temperatures, residence time, and acidity). High values (0 to 3) represent harsher conditions (i.e. relatively high temperatures, residence time, and acidity). Second, the practical temperature operating range is between 100 and 230 o C. Temperatures exceeding 230 o C will drive significant thermal degradation of all polysaccharides and monosaccharides, leaving behind mostly lignin in the product (which is not usable for microbial digestion). Third, since the CSF equation is based on the Arrhenius equation for acid catalysis, liquor pH of 7 or less can only be used. (Chum, 1990)
1.3.2 Acid hydrolysis
There are numerous reactions that take place in aqueous sulfuric and other strong acid media. This includes hydrolyses, dehydrations, hydrations, isomerizations, electrophilic substitutions, aromatic rearrangements, carbonyl reactions, and a number of other reactions. (Cox, 1987)
Sulfuric acid has also been added to cellulosic materials for many years, particularly in the pulp-and-paper manufacturing bleaching process (Root et al., 1959; Zeitsch, 2000). This acid has been widely used and studied for pretreatment. In this work, sulfuric acid was used to catalyze the hydrolysis of polysaccharides found in biomass.
The molecular mechanism of acid-catalyzed cellulose hydrolysis is represented by the cleavage of the β-1-4-glycosidic bond (Xiang, 2003). This is a homogeneous reaction in which the acid catalyzes the breakdown of cellulose to produce oligomers (cellobiose) and monosaccharides (glucose). The rate of thermal induced degradation is accelerated in the presence of water, acids and oxygen. As the temperature increase, the degree of polymerization of cellulose decreases further, free radicals appear and carbonyl, carboxyl and hydroperoxide groups are formed. This undesirable and independent reaction involves the breakdown of glucose to form degradation products, such as xylitol, succinic acid, L-lactic acid, glycerol, acetic acid, ethanol, 5-hydroxy-2-furaldehyde, and furfural
(Hu, 2008). Excessively severe conditions such as high acid loading or high temperatures can result in oxidative degradation of carbohydrates, yielding fermentation inhibitors (Mosier, 2003).
Kinetic modeling plays a key role in the design, development, and operation of reactors. Kinetic data are also vital in the design and evaluation of processes to hydrolyze cellulosic materials to glucose for ethanol conversion (Conner, 1985).