Bioresource Technology 98 (2007) 3000–3011

A comparison of chemical pretreatment methods for improving saccharification of cotton stalks Rebecca A. Silverstein a, Ye Chen a, Ratna R. Sharma-Shivappa Michael D. Boyette a, Jason Osborne b a

a,*

,

Department of Biological and Agricultural Engineering, Campus Box 7625, North Carolina State University, Raleigh, NC 27695-7625, USA b Department of Statistics, Campus Box 8203, North Carolina State University, Raleigh, NC 27695-7625, USA Received 10 October 2006; accepted 12 October 2006 Available online 8 December 2006

Abstract The effectiveness of sulfuric acid (H2SO4), sodium hydroxide (NaOH), hydrogen peroxide (H2O2), and ozone pretreatments for conversion of cotton stalks to ethanol was investigated. Ground cotton stalks at a solid loading of 10% (w/v) were pretreated with H2SO4, NaOH, and H2O2 at concentrations of 0.5%, 1%, and 2% (w/v). Treatment temperatures of 90 C and 121 C at 15 psi were investigated for residence times of 30, 60, and 90 min. Ozone pretreatment was performed at 4 C with constant sparging of stalks in water. Solids from H2SO4, NaOH, and H2O2 pretreatments (at 2%, 60 min, 121 C/15 psi) showed significant lignin degradation and/or high sugar availability and hence were hydrolyzed by Celluclast 1.5 L and Novozym 188 at 50 C. Sulfuric acid pretreatment resulted in the highest xylan reduction (95.23% for 2% acid, 90 min, 121 C/15 psi) but the lowest cellulose to glucose conversion during hydrolysis (23.85%). Sodium hydroxide pretreatment resulted in the highest level of delignification (65.63% for 2% NaOH, 90 min, 121 C/15 psi) and cellulose conversion (60.8%). Hydrogen peroxide pretreatment resulted in significantly lower (p 6 0.05) delignification (maximum of 29.51% for 2%, 30 min, 121 C/15 psi) and cellulose conversion (49.8%) than sodium hydroxide pretreatment, but had a higher (p 6 0.05) cellulose conversion than sulfuric acid pretreatment. Ozone did not cause any significant changes in lignin, xylan, or glucan contents over time. Quadratic models using time, temperature, and concentration as continuous variables were developed to predict xylan and lignin reduction, respectively for sulfuric acid and sodium hydroxide pretreatments. In addition, a modified severity parameter (log M0) was constructed and explained most of the variation in xylan or lignin reduction through simple linear regressions.  2006 Elsevier Ltd. All rights reserved. Keywords: Delignification; Bioethanol; Modeling; Lignocellulose; Enzymatic hydrolysis

1. Introduction Growing concerns over the environmental impact of fossil fuels and their inevitable depletion have led to intense research on the development of alternative energy sources that can reduce the United States dependence on foreign oil imports. Biomass, which includes animal and human waste, trees, shrubs, yard waste, wood products, grasses, and agricultural residues such as wheat straw, corn stover,

*

Corresponding author. Tel.: +1 919 515 6746; fax: +1 919 515 7760. E-mail address: [email protected] (R.R. Sharma-Shivappa).

0960-8524/$ - see front matter  2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2006.10.022

rice straw, and cotton stalks, is a renewable resource that stores energy from sunlight in its chemical bonds (McKendry, 2002). It can be processed either chemically or biologically by breaking the chemical bonds to extract energy in the form of biofuels such as bioethanol, biodiesel, and methane. Currently, corn is the primary raw material for ethanol production in the United States. Starch, which constitutes about 70% of the corn kernel, is easily broken down into glucose that is then fermented to ethanol. However, the corn to ethanol industry draws its feedstock from a food stream and is quite mature with little possibility of process improvements (Ingram and Doran, 1995). Lignocellulosic

R.A. Silverstein et al. / Bioresource Technology 98 (2007) 3000–3011

feedstocks, which have the potential to reduce the cost of producing ethanol because they are less expensive than corn and available in large quantities, offer a plausible alternative. One promising technology is to convert this abundant and renewable lignocellulosic biomass to ethanol through an enzyme-based process (Schell et al., 2003). The conversion of lignocellulosic biomass to ethanol is, however, more challenging than corn due to the complex structure of the plant cell wall. Pretreatment is required to alter the structural and chemical composition of lignocellulosic biomass to facilitate rapid and efficient hydrolysis of carbohydrates to fermentable sugars (Chang and Holtzapple, 2000). A variety of physical (comminution, hydrothermolysis), chemical (acid, alkali, solvents, ozone), physico-chemical (steam explosion, ammonia fiber explosion), and biological pretreatment techniques have been developed to improve the accessibility of enzymes to cellulosic fibers (Moiser et al., 2005). Acid pretreatment involves the use of sulfuric, nitric, or hydrochloric acids to remove hemicellulose components and expose cellulose for enzymatic digestion (Schell et al., 2003). Agricultural residues such as corncobs and stovers have been found to be particularly well suited to dilute acid pretreatment (Torget et al., 1991). Alkali pretreatment refers to the application of alkaline solutions to remove lignin and various uronic acid substitutions on hemicellulose that lower the accessibility of enzyme to the hemicellulose and cellulose (Chang and Holtzapple, 2000). Generally, alkaline pretreatment is more effective on agricultural residues and herbaceous crops than on wood materials (Hsu, 1996). Peroxide pretreatment enhances enzymatic conversion through oxidative delignification and reduction of cellulose crystallinity (Gould, 1985). Increased lignin solubilization and cellulose availability were observed during the peroxide pretreatment of wheat straw (Martel and Gould, 1990), Douglas fir (Yang et al., 2002), and oak (Kim et al., 2001). Ozonation is another attractive pretreatment method that does not leave strong acidic, basic, or toxic residues in the treated material (Neely, 1984). The effect of ozone pretreatment has been found to be essentially limited to lignin degradation. Hemicellulose is slightly attacked, while cellulose is hardly affected (Sun and Cheng, 2002). Ozonation has been widely used to reduce the lignin content of both agricultural and forestry wastes (Neely, 1984). Cotton (Gossypium hirsutum), which is one of the most abundant crops in the southern United States, apart from being invaluable for the textile industry is a significant source of lignocellulosic biomass. In 2003, nearly 13.2 million acres of cotton were planted nationwide as a result of increased world demand for cotton (USDA, 2004). The increase in cotton planting is highly beneficial for economic development, but it also raises concerns about the disposal of cotton stalks left in the field (TBWEF, 2004) that serve as breeding ground for pests. Cotton stalks, which mainly contain lignocellulose, have the potential to serve as a low-cost feedstock to increase the production of fuel

3001

ethanol through proper pretreatment, hydrolysis, and fermentation. Conversion of this agricultural waste into a value-added product can provide an environmentally sound method of disposal and simultaneous destruction of feeding and fruiting sites of boll weevils and other insects. To fully utilize cotton stalk as a feedstock for ethanol production, optimal pretreatment is required to render the cellulose fibers more amenable to the action of hydrolytic enzymes. This study was therefore initiated to: (1) investigate the effect of sulfuric acid, sodium hydroxide, hydrogen peroxide, and ozone pretreatments of cotton stalks, (2) develop models to predict lignin degradation and xylan solubilization percentage during sulfuric acid and sodium hydroxide pretreatments, and (3) identify pretreatment(s) which provide the highest cellulose to glucose conversion during subsequent enzymatic hydrolysis. 2. Methods 2.1. Biomass Cotton stalks, harvested in early October 2003, were obtained from Cunningham Research Station in Kinston, NC. The stalks were shredded and bailed in the field soon after the cotton was picked, and then transported to North Carolina State University in Raleigh, NC. Prior to composition analysis, the biomass which consisted primarily of stalks, leaves, cottonseed, and cotton residue, was ground to a 40 mesh particle size. The feedstock was ground to pass a 3 mm sieve in a Thomas Wiley Laboratory Mill (Model No. 4) and stored in sealed plastic bags at room temperature until use for pretreatment. 2.2. Analysis methods The total solids, acid soluble lignin, and acid insoluble lignin (acid-insoluble material) content of the untreated cotton stalks and the solid fraction remaining after pretreatment were determined by Laboratory Analytical Procedures (LAP) from the National Renewable Energy Laboratory (NREL) (Ehrman, 1994, 1996; Templeton and Ehrman, 1994). Ash content, extractives, and holocellulose (combination of hemicellulose and cellulose), were determined for the untreated stalks by the gravimetric methods developed by Han and Rowell (1997). The carbohydrate contents of the untreated cotton stalks and pretreated solids were determined by measuring the hemicellulose (xylan, galactan, and arabinan) and cellulose (glucan) derived sugars. The composition of the hydrolysate from enzymatic hydrolysis was determined by measuring glucose and xylose using high performance liquid chromatography (HPLC). The LAP-002 analysis procedure from NREL was modified for use with a Dionex-300 HPLC system (Ruiz and Ehrman, 1996). The HPLC system was equipped with a CarboPactrade PA10 (4 · 250 mm) anion exchange column, a guard column

3002

R.A. Silverstein et al. / Bioresource Technology 98 (2007) 3000–3011

(4 · 50 mm), an automated sampler, a gradient pump, and a pulsed amperometric detector with a gold working electrode (Dionex Corp., Sunnyvale, CA). The mobile phase used was 10 mM NaOH at a flow rate of 1 mL/min. Monomeric sugars (arabinose, galactose, glucose, xylose) with the concentrations of 0, 10, 30, and 50 mg/l were used as standards. Prior to HPLC injection, all samples (derived from solids and hydrolysate) were neutralized with barium hydroxide, centrifuged at 5000g for 10 min, and filtered through 0.45 lm Millipore filters. Fucose was added as an internal standard for the samples analyzed.

at 5000g for 10 min. The supernatant was removed for sugar content analysis (Yang and Wyman, 2004). The percent glucan conversion was calculated as follows: % glucan conversion ¼

%GH  100 %GP

ð1Þ

where GH is the dry-weight percentage of glucose in enzyme hydrolysis supernatant (g glucose/g solids hydrolyzed %), GP is the dry-weight percentage glucan in pretreated solids (g glucose/g solids pretreated %). A similar equation was used to determine percent xylan conversion.

2.3. Pretreatments Sulfuric acid (H2SO4), sodium hydroxide (NaOH), and hydrogen peroxide (H2O2) at concentrations of 0.5%, 1%, and 2% (w/v) were used to pretreat 10 g ground cotton stalk samples at a solid loading of 10% (w/v). Treatments were performed in triplicate at 90 C and in an autoclave at 121 C with 15 psi (103.4 kPa) pressure for residence times of 30, 60, and 90 min. Ozone pretreatment was performed by continuously sparging ozone gas through a 10% (w/v) mixture of cotton stalks and deionized water for 30, 60, and 90 min. The reaction temperature was controlled at 4 C by placing the reaction flask in a water bath. Ozone gas was generated on site by passing 5 L/min of oxygen through an ozonator (AOS-1M/MS, Applied Ozone Systems, CA). Ozone concentrations in pure deionized water sparged with ozone were determined by measuring the absorbance using a Shimadzu PharmaSpec UV-1700 spectrophotometer (Columbia, MD) at 258 nm (Sharma et al., 2002). The pretreated solids were washed with 750 mL of hot deionized water and used for determination of total solids, acid insoluble lignin, and glucan and xylan prior to storage at 4 C for enzymatic hydrolysis. 2.4. Enzymatic hydrolysis Cellulase from Trichoderma reesei (Celluclast 1.5 L, Sigma Co., St. Louis, MO) with an activity of 96.1 filter paper unit FPU/mL enzyme solution, supplemented with cellobiase from Aspergillus niger (Novozyme 188, EC No. 232-589-7, Sigma Co., St. Louis, MO) at a ratio of 1:1.75 was used for hydrolysis experiments. The protein contents of Celluclast 1.5 L and Novozym 188 have been reported to be 191 and 143 mg/mL, respectively (Shoemaker, 2004). Pretreated cotton stalks at 5% solids loading (grams dry weight per 100 mL) in 50 mM acetate buffer (pH 4.8) containing 40 lg/mL tetracycline (an antibiotic added to avoid microbial contamination) were preincubated in flasks in a shaking water bath at 50 C and 150 rpm for 10 min. The hydrolysis, conducted at a cellulase activity of 40 FPU/g cellulose, was initiated by adding 2.18 mL and 3.82 mL of cellulase and cellobiase, respectively. Aliquots of 2.0 mL were taken at the termination of enzymatic hydrolysis after 72 h, immediately chilled on ice, and centrifuged

2.5. Data analysis and modeling The experimental design was crossed and complete with respect to temperature, time and concentration. Factorial effects models with main and interaction effects, on lignin reduction and xylan and glucan solubilization by sulfuric acid, sodium hydroxide, hydrogen peroxide, and ozone pretreatments, were fit using PROC GLM in SAS (SAS Institute, Cary, NC). Multiple comparisons among treatment means were carried out using Tukey’s procedure to control the experiment wise error rate at 0.05 for each response variable. In cases where interaction effects were significant, the SLICE option was used with the LSMEANS command of the GLM procedure to test for simple treatment effects of one factor while holding the other two factors constant. Empirical quadratic models with time, temperature, and concentration as continuous numeric variables were developed using SAS to predict percent lignin reduction for sodium hydroxide pretreatment and xylan solubilization for sulfuric acid pretreatment and were of the form y ¼ b0 þ b1 T þ b2 t þ b3 C þ b4 Tt þ b5 CT þ b6 CTt þ b7 Ct þ b8 t2 þ b9 C 2

ð2Þ

where T is the temperature (C); t is the time (min); C, the concentration (%); bn, the estimated regression coefficients, n = 0, 1, . . . , 9. The squared temperature term (T2) was not included in the model because only two temperatures were used during the experiments. This did not provide a sufficient number of degrees of freedom to estimate a regression coefficient for a squared term. Additionally, since the data set was relatively small to quantify the predictability of the models, the focus was on model development rather than assessment of predictive ability of the models. In addition, modeling based on combining the effects of time, temperature, and concentration into one single parameter was used to develop a linear model expressing the relationship between pretreatment severity and lignin reduction or xylan solubilization. Overend and Chornet (1987) initially defined this severity parameter to relate temperature and time for steam explosion pretreatment based on the assumption that pretreatment affects follow

R.A. Silverstein et al. / Bioresource Technology 98 (2007) 3000–3011

first-order kinetics and obey the Arrhenius equation. Using this relationship they defined a reaction ordinate (R0, min)   ðT r  T b Þ R0 ¼ t  exp ð3Þ 14:75 where t is the residence time (min), Tr is the reaction temperature (C), Tb is the base temperature (100 C) and 14.75 is the conventional energy of activation assuming the overall reaction is hydrolytic and the overall conversion is first order. The logarithm of the reaction ordinate defines the severity during steam explosion pretreatment such that severity is equal to log (R0). A modified severity parameter was later developed by Chum et al. (1990) for use with sulfuric acid pretreatment   Tr  Tb M 0 ¼ t  C n  exp ð4Þ 14:75 where M0 is the modified severity parameter; t is the residence time (min); C is the chemical concentration (wt%); Tr is the reaction temperature (C); Tb is the base temperature (100 C); n is an arbitrary constant. This equation was adapted for application to sodium hydroxide pretreatment by replacing the acid concentration with the alkali concentration and calculating a different n-value. 3. Results and discussion 3.1. Characterization of cotton stalk The chemical composition of cotton stalks varies depending on the growing location, season, harvesting methods, as well as analysis procedures (Agblevor et al., 2003). The composition of the cotton stalks used in this study is presented in Table 1. Based on the HPLC carbohydrate analysis, the sugar fraction was 41.8% of the dry biomass. Glucan, which is derived from both the cotton fiber and plant cell wall, was the major component at 31.1%. Xylan, which is the major hemicellulose constituent, was 8.3%. Arabinan and galactan accounted for only a small portion of the biomass, while mannan was not detected.

Table 1 Summative composition of untreated cotton stalks Component

Percentage (%)a

Holocellulose Glucan Xylan Arabinan Galactan Acid-insoluble lignin Acid-soluble lignin Extractives Ash Other

41.8 31.1 8.3 1.3 1.1 27.9 2.2 9.0 6.0 13.1

a

Composition percentages are on a dry-weight basis.

3003

Both glucan and xylan content were lower than the reported values of 40–50% glucan and 15–35% xylan for other agricultural residues and hardwoods (Milne et al., 1992). The holocellulose fraction, determined by the procedure of Han and Rowell (1997), was 51.1% of the total biomass. The discrepancy between the holocellulose content and total sugars is probably due to sugar degradation during the intense hydrolysis with sulfuric acid (Badger, 2002) used for the carbohydrate analysis procedure. Since the carbohydrate content of cotton stalks (based on monomeric sugars) determined by HPLC analysis was more likely to represent the actual sugars available after the treatments, subsequent calculations and analysis of data in this study were performed on the basis of HPLC measurements. The acid-insoluble material content of the cotton stalks (27.9%) was higher than expected. It was comparable to the acid-insoluble material content of hardwoods (18–25%), rather than that of herbaceous species and agricultural residues (10–20%) (McMillan, 1994). The acid insoluble material from woody biomass is normally classified as lignin. However, it would be incorrect to classify all the acid insoluble material from cotton stalks as lignin. A possible source of non-lignin, acid-insoluble material is the cottonseed. Cottonseed is composed of 32% hull, 23% protein, 12% fibers, 20% oil, and 14% carbohydrates. Upon analysis of the cottonseed from the Emporia gin in Virginia, Agblevor et al. (2003) discovered that the cottonseed contained 34% acid-insoluble material. The hull, which is lignocellulosic, and thus the only source of lignin, makes up only 32% of the cottonseed. Thus, the acid-insoluble material is expected to be composed of lignin and other condensable compounds. Since it is known that proteins condense and become insoluble in concentrated sulfuric acid (Agblevor et al., 1994), it could be surmised that high acid-insoluble material of the cottonseed, and in turn, the cotton stalks, is a combination of lignin and condensed proteins (Agblevor et al., 2003). However, because the majority of the acid insoluble material is lignin, it has been referred to as such in this study to limit confusion. The composition of cotton stalks used in this study agreed with that of cotton gin residue (immature bolls, cottonseed, hulls, sticks, leaves, and dirt) analyzed by Agblevor et al. (2003). The residue was sampled two to three times on different days from five different cotton gins across Virginia. The composition varied depending on the discharge date and the gin location, with approximate ranges of each component being 21–38% glucan, 3–12% xylan, 0.5–3% each of mannan, galactan, and arabinan, 5–13% extractives, 18–26% acid-insoluble lignin, and 7–14% ash. 3.2. Effect of sulfuric acid pretreatment Dilute acid pretreatment of lignocellulosic biomass is one of the most effective pretreatment methods which predominantly affect hemicellulose with little impact on lignin degradation. The lignin reduction, xylan and glucan

3004

R.A. Silverstein et al. / Bioresource Technology 98 (2007) 3000–3011

solubilization during sulfuric acid pretreatment of cotton stalks is shown, respectively in Fig. 1a–c and the solids recovery after pretreatment is presented in Table 2. ANOVA tables providing information on which treatment parameters had significant impact during pretreatment are not presented in this text but are available from Silverstein (2004). Elevated temperature, residence time, and acid concentration reduced solids recovered after pretreatment. The acid insoluble lignin, remaining after acid pretreatment varied from 28.72% (1%, 30 min, 90 C) to 40.68% (2%, 60 min, 121 C/15 psi). The xylan and glucan contents of the sulfuric acid pretreated samples ranged from 0% (2%, 60 min, 121 C/15 psi) to 10.24% (0.5%, 30 min, 90 C) and 33.74% (2%, 30 min, 90 C) to 46.3% (2%, 60 min, 121 C/15 psi) respectively. The reduction of lignin, based on a comparison between the weight of lignin in the initial

10 g (dry-weight) sample before pretreatment and the weight of lignin in the solids remaining after pretreatment, ranged from 2.27% to 24.16%. Concentration had a significant (p 6 0.05) effect on delignification for treatments at 90 C for 60 min and 121 C/15 psi for 30, 60, and 90 min. Increasing the temperature from 90 C to 121 C/ 15 psi significantly increased delignification with pretreatment for 60 min at 2% H2SO4 and 90 min at 0.5%, 1%, and 2% H2SO4. The filtrate from the lignin analysis was used to quantify the remaining hemicellulose and cellulose. The xylan content, which makes up the largest portion of hemicellulose in the cotton stalks, is the most important indicator of pretreatment effectiveness. Arabinan and galactan, although making up 1.3% and 1.1%, respectively of the untreated sample, were below the HPLC detection limit after

Lignin reduction (%)

35 30 25 20 15 10 5 0 30

60

90

30

90

60

90

Time (min) Temp. (o C)

121/15 psi

Xylan solubilization (%)

120 100 80 60 40 20 0 30

60

90

30

90

60

90

Time (min) Temp. (o C)

121/15 psi

Glucan solubilization (%)

30 25 20 15 10 5 0 30

60 90

90

30

60 121/15 psi

90

Time (min) Temp. (o C)

Fig. 1. (a) Lignin reduction, (b) xylan solubilization, and (c) glucan solubilization in sulfuric acid pretreated samples as a function of residence time, temperature, and concentration.

R.A. Silverstein et al. / Bioresource Technology 98 (2007) 3000–3011 Table 2 Percent solids recovery after pretreatments Time (min), conc. (%), temp. (C)

Solids recovered after pretreatment (%)a,b Sulfuric acid

Sodium hydroxide

Hydrogen peroxide

30, 30, 30, 60, 60, 60, 90, 90, 90, 30, 30, 30, 60, 60, 60, 90, 90, 90,

80.27 85.05 81.18 83.32 83.56 73.42 83.83 77.87 75.02 75.45 67.71 62.16 73.07 64.44 56.93 68.23 60.56 56.95

75.47 70.42 62.22 70.79 60.64 53.40 74.17 63.57 55.39 71.56 68.24 60.02 69.16 59.07 55.14 72.95 58.11 54.50

83.16 82.27 83.51 86.02 84.28 81.93 85.59 83.21 80.44 84.12 81.64 76.83 83.97 81.15 74.42 85.04 79.26 72.59

0.5, 1.0, 2.0, 0.5, 1.0, 2.0, 0.5, 1.0, 2.0, 0.5, 1.0, 2.0, 0.5, 1.0, 2.0, 0.5, 1.0, 2.0,

90 90 90 90 90 90 90 90 90 121/15 psi 121/15 psi 121/15 psi 121/15 psi 121/15 psi 121/15 psi 121/15 psi 121/15 psi 121/15 psi

RMSE R-square Tukey’s HSD

2.59 0.95 8.04

(3.86) (2.81) (4.24) (2.61) (0.36) (3.28) (1.61) (4.96) (2.56) (1.12) (1.97) (0.41) (0.73) (1.88) (3.04) (3.04) (1.98) (0.99)

3.22 0.88 10.02

(3.92) (1.28) (3.30) (4.58) (2.50) (1.69) (1.44) (1.38) (3.98) (7.97) (4.87) (2.03) (0.83) (1.49) (1.29) (1.32) (1.82) (1.02)

(1.57) (1.41) (0.52) (1.37) (1.61) (2.21) (5.83) (1.39) (2.43) (1.36) (0.81) (0.82) (2.19) (4.83) (2.90) (1.55) (1.76) (0.60)

2.38 0.78 7.28

a

Percentages calculated from value on a dry-weight basis. Data are averages of three replicates. Numbers in parentheses represent standard deviations. b

pretreatment, and therefore xylan was the only hemicellulose sugar determined hereafter. Sulfuric acid pretreatment effectively solubilized 14.57% of the xylan for the least severe pretreatment (0.5%, 30 min, 90 C) and 95.2% for the most severe treatment (2%, 90 min, 121 C/15 psi) (Fig. 1b). Results from this study are comparable to those obtained by Varga et al. (2002), who observed 85% solubilization of hemicellulose in corn stover at 121 C/15 psi for 1 h with 5% H2SO4. Increasing temperature had most pronounced effect on xylan solubilization. The temperature effect was significant for all combinations of time and concentration, showing that 121 C/15 psi is more effective for xylan solubilization than 90 C. No significant concentration effect on xylan solubilization was detected at 90 C for 30 min or 60 min treatment (p > 0.05). This indicated increasing acid concentration for the two lowest combinations of time and temperature did not increase the amount of xylan solubilization. In addition, there was no significant time effect at 90 C, 0.5% acid, indicating that the severity of the treatment at the lowest concentration and temperature does not show any significant improvement with an increase in time from 30 to 90 min. At 2% acid, 60 min, and 121 C/15 psi no xylose was detected. The results obtained at 2% acid, 90 min, and 121 C/15 psi were similar, with two of the three replicates showing no detectable levels of xylan and the third sample possessing only 2.44% xylose. Possible explanations for this could be that (1) there was complete solubilization of xylan during pretreatment and/or (2) the amount of xylan remaining in the sample,

3005

treated at 121 C/15 psi for 60 and 90 min and used for analysis was lower than the detection limit for HPLC analysis. The percentage of glucan solubilization due to sulfuric acid pretreatments was between 10.00% (0.5%, 60 min, 121 C/15 psi) and 23.88% (2%, 90 min, 121 C/15 psi) (Fig. 1c). Temperature and concentration had significant (p 6 0.05) effect on glucan solubilization. During pretreatment it is desirable that the cellulose portion of the biomass remain virtually unaffected. However, during acid pretreatment of cotton stalks slightly higher glucan reduction was observed (Kim et al., 2001) because the cellulose rich, loose cotton fiber, in the stalks, is not imbedded in lignin and hemicellulose. The acid therefore has direct access to the cellulose during pretreatment and can cause more glucan degradation than is usually the case with other feedstocks. 3.3. Effect of sodium hydroxide pretreatment Using sodium hydroxide to pretreat lignocellulosic materials is an alternative to sulfuric acid pretreatment. The main effect of sodium hydroxide pretreatment on lignocellulosic biomass is delignification by breaking the ester bonds cross-linking lignin and xylan, thus increasing the porosity of biomass (Tarkov and Feist, 1969). As in acid pretreatment, elevation in temperature, residence time, and alkali concentration increased the loss of solids during NaOH pretreatment (Table 2). The amount of lignin in the solids after NaOH pretreatment ranged from 23.31% (30 min, 90 C) to 25.22% (30 min, 121 C/ 15 psi) for 0.5% NaOH, 19.46% (60 min, 121 C/15 psi) to 21.90% (30 min, 90 C) for 1% NaOH, and 17.64% (90 min, 121 C/15 psi) to 20.94% (30 min, 121 C/15 psi) for 2% NaOH. Changes in concentration caused significant (p 6 0.05) decrease in lignin. The xylan content of pretreated solids ranged from 7.91% (0.5%, 30 min, 90 C) to 13.00% (1%, 90 min, 121 C/15 psi) and the glucan content ranged from 35.54% (0.5%, 30 min, 90 C) to 50.33% (2%, 60 min, 121 C/15 psi). The maximum reduction in lignin was 65.63% with 2% NaOH treatment for 90 min at 121 C/15 psi (Fig. 2a). Varga et al. (2002) reported 95% reduction in lignin content as a result of pretreatment of corn stover with 10% NaOH for 1 h in the autoclave. The high reduction level may be attributed to a higher NaOH concentration of 10%, which in this study was limited to 2%. An increase in the concentration of NaOH significantly improved delignification at all combinations of temperature and time (p 6 0.05). There was no significant (p > 0.05) effect of time with 0.5% NaOH at either temperature, indicating that 0.5% NaOH is too low to affect delignification for treatment times up to 90 min and temperatures up to 121 C in the autoclave. The effect of temperature for sodium hydroxide pretreatment was significant (p 6 0.05) only when the residence time was 90 min at 1% and 2% NaOH. This indicates that increasing the temperature only improved the amount of lignin removal for longer times and higher concentrations.

3006

R.A. Silverstein et al. / Bioresource Technology 98 (2007) 3000–3011

Lignin reduction (%)

70 60 50 40 30 20 10 0 30

60

90

30

90

60

90

Time (min) Temp. (oC)

121/15 psi

Xylan solubilization (%)

80 60 40 20 0 30

60

90

30

Glucan solubiliza tion (%)

90

60

90

Time (min) Temp. (o C)

121/15 psi

50 40 30 20 10 0 30

60

90

90

30

60 121/15 psi

90

Time (min) Temp. (o C)

Fig. 2. (a) Lignin reduction, (b) xylan solubilization, and (c) glucan solubilization in sodium hydroxide pretreated samples as a function of residence time, temperature, and concentration.

Although xylan solubilization due to sodium hydroxide pretreatment was lower than that by sulfuric acid pretreatment (Fig. 2b), it is expected that solubilization of xylan in conjunction with substantial lignin reduction can improve enzymatic hydrolysis. Sodium hydroxide pretreatment resulted in xylan solubilization in the range of 13.90% (0.5%, 90 min, 90 C) to 40.02% (2%, 90 min, 90 C). Concentration, time, and temperature did not cause significant (p > 0.05) differences in percent xylan solubilization for any of the treatment combinations. The solubilization of glucan during NaOH pretreatment was between 12.82% (1%, 30 min, 121 C/15 psi) and 30.14% (2%, 60 min, 90 C) as illustrated in Fig. 2c. Glucan solubilization increased significantly with increasing concentration for 90 C at 90 min and the temperature effect was significant for 2% NaOH for 30 and 60 min. However,

the standard deviations among some replicates were rather large. This could be attributed to the heterogeneous nature of cotton stalks and the fact that amount of free cotton fiber could vary from one sample to the other.

3.4. Effect of hydrogen peroxide pretreatment Hydrogen peroxide pretreatment utilizes oxidative delignification to detach and solubilize the lignin and loosens the lignocellulosic matrix thus improving enzyme digestibility (Martel and Gould, 1990). The lignin reduction and xylan and glucan solubilization due to H2O2 pretreatment in this study are shown in Fig. 3a–c and the solids recovered after pretreatment are presented in Table 2. There was no evidence (p > 0.05) of any effects of either

R.A. Silverstein et al. / Bioresource Technology 98 (2007) 3000–3011

3007

Lignin reduction (%)

40 30 20 10 0 30

60

90

30

90

60

90

Time (min) Temp. (o C)

121/15 psi

Xylan solubilization (%)

40 30 20 10 0 30

60

90

30

90

60

90

Time (m in) Te m p. (o C)

121/15 psi

Glucan solubilization (%)

40 30 20 10 0 30

60

90

90

30

60 121/15 psi

90

Time (min) Temp. (o C)

Fig. 3. (a) Lignin reduction, (b) xylan solubilization, and (c) glucan solubilization in hydrogen peroxide pretreated samples as a function of residence time, temperature, and concentration.

of the treatment factors on the lignin content of pretreated solids. Hydrogen peroxide pretreatment led to 6.22% (0.5%, 90 min, 90 C) to 32.01% (2%, 60 min, 121 C/15 psi) delignification (Fig. 3a). These lignin degradations are lower than those reported in literature at alkaline conditions where pretreatment of sugar cane bagasse with 2% alkaline H2O2 resulted in 50% decrease in lignin and solubilization of most of the hemicellulose within 8 h at 30 C (Azzam, 1989). Determination of simple treatment effects for delignification showed that increasing the concentration from 0.5% to 2% did not significantly increase delignification for 30 min at 90 C probably because the residence time was too short at the lower temperature. The simple time effect was significant for 121 C/15 psi at 0.5% and 1% H2O2, which indicates that increasing the residence time

from 30 to 90 min showed significant improvements only for the two lower concentrations at the higher temperature. Temperature played a significant role in improving delignification for 0.5% at 60 min and 2% at 30 and 60 min but an increase in temperature significantly reduced the mean delignification for 0.5% at 90 min. The most severe pretreatment in the autoclave at 121 C for 90 min with 2% H2O2 had lower levels of delignification than the treatments at 30 and 60 min at 0.5% and 1%. This could be attributed to the decomposition of H2O2 at high temperature thus diminishing its oxidative delignification potential and to the long residence time which could result in recondensation or repolymerization of solubilized lignin. The solubilization of xylan due to H2O2 pretreatment averaged between 8.18% (0.5%, 60 min, 90 C) and 30.56% (2%, 30, 121 C/15 psi) (Fig. 3b) while the xylan

3008

R.A. Silverstein et al. / Bioresource Technology 98 (2007) 3000–3011

content ranged from 8.69% (2%, 30 min, 121 C/15 psi) to 10.87% (2%, 90 min, 90 C). Concentration had a significant effect (p 6 0.05) on xylan solubilization for 90 min, 90 C treated samples and 30 min, 121 C treated samples. The simple temperature effect was significant for xylan solubilization for 0.5% and 1% at 60 min and 2% at 30, 60, and 90 min. The percentage of glucan in the pretreated solids remaining as a result of H2O2 pretreatment ranged from 28.4% (1%, 30 min, 90 C) to 34.1% (2%, 90 min, 90 C). Glucan solubilization on average was between 14.91% (0.5%, 60 min, 90 C) to 29.10% (2%, 30 min, 121 C/ 15 psi) as presented in Fig. 3c. Concentration did not have a significant effect on glucan solubilization. Significant differences (p 6 0.05) between percent glucan solubilization due to changes in temperature were noted for 2% at 90 min, while time played a significant role for 0.5 and 1% H2O2 at 90 C. 3.5. Effect of ozone pretreatment Pretreatment of lignocellulosic biomass with ozone gas has been reported to reduce both the lignin and hemicellulose contents of the treated materials (Ben-Ghedalia et al., 1980). The most substantial effect of ozone pretreatment is on degradation of the lignin polymer, followed by hemicellulose and cellulose solubilization (Quesada et al., 1999). In this study, ozone pretreatment reduced lignin in the range of 11.97–16.6% with no significant difference (p > 0.05) noted for treatment times of 30, 60, and 90 min (Table 3). The amount of xylan solubilized during ozone treatment ranged from 1.9% to 16.7%, while the amount of glucan solubilized was between 7.2% and 16.6%. The percent solubilization of xylan and glucan for 90 min treatment was significantly (p < 0.05) lower than the solubilization for 30 and 60 min. The concentrations of ozone measured in pure water after sparging for 30, 60, and 90 min were 16.96, 17.74, and 18.52 ppm, respectively. Ben-Ghedalia et al. (1980) reported a 50% decrease in both lignin and hemicellulose in ozone treated cotton stalks. Possible explanations for the differences between the results from this study and those from past studies include insufficient treatment times, inadequate ozone concentration, and

poor distribution of ozone gas throughout the cotton stalks because of inefficient sparging. 3.6. Enzymatic hydrolysis Acid pretreated samples resulting in maximum glucose availability (2% H2SO4, 60 min, 121 C/15 psi) were chosen for enzyme hydrolysis. This selection criterion was based on the fact that acid pretreatment has little effect on lignin degradation and the main treatment effect is on hemicellulose and cellulose solubilization. Alkali pretreatment caused delignification and glucan solubilization. The selection for NaOH pretreated samples was based on a compromise between having the lowest percentage of lignin in the pretreated solids, while maintaining a high percentage of glucan (2% NaOH, 60 min, 121 C/15 psi). For hydrogen peroxide, there were no significant differences between percentage glucan, xylan, or lignin in the pretreated solids for any of the treatments. Hence, the treatment with the highest percentage of glucan and the lowest percentage of lignin was chosen (2% H2O2, 60 min, 121 C/15 psi). Cellulose conversion of pretreated samples after 72 h of enzymatic hydrolysis is shown in Table 4. Sodium hydroxide pretreated samples had the highest cellulose conversion of 60.8%, followed by hydrogen peroxide (49.8%) and then sulfuric acid (23.8%). Differences in mean cellulose conversions for all the treatments were statistically significant (p 6 0.05). Hydrolysis of sodium hydroxide pretreated samples resulted in the highest xylan to xylose conversion (Table 4) at 62.57%, whereas hydrogen peroxide averaged 7.78% conversion. For the acid pretreated samples, no xylan was detected in the solids during the initial carbohydrate analysis, but an average of 14.3 mg xylose/g dry biomass was detected in the supernatant after enzymatic hydrolysis. This confirms the hypothesis that there was xylan in the stalks after pretreatment, but the amount was below the detection limit during sugar analysis. Detection of xylose in the hydrolysate may be attributed to a higher sugar concentration resulting from the hydrolyzed sample (5 g) being larger than that analyzed for carbohydrate content of pretreated solids (0.3 g). The difference in cellulose conversion during enzymatic hydrolysis is largely dependent on the difference in lignin

Table 3 Effect of ozone pretreatment on cotton stalks Time (min)

30 60 90c RMSE R-square Tukey’s HSD a b c

Reduction (%)a,b

Solids recovery (%)

Lignin

Xylan

Glucan

11.97 (2.91) 16.63 (2.60) 15.15 (3.02)

16.76 (7.32) 10.61 (8.12) 1.92 (2.89)

16.62 (7.80) 13.74 (3.64) 7.19 (0.36)

7.04 0.52 20.19

5.45 0.42 15.63

2.82 0.46 8.08

Percentages calculated from values on a dry-weight basis. Data are averages of three replicates. Numbers in parentheses represent standard deviations. Only two samples were used for 90 min treatment because the third replicate was an outlier.

90.44 (2.47) 88.66 (2.82) 91.66 (0.47) 2.18 0.32 5.46

R.A. Silverstein et al. / Bioresource Technology 98 (2007) 3000–3011

3009

Table 4 Glucan and xylan conversion after enzymatic hydrolysis Pretreatment agent

Composition of hydrolysis supernatant and pretreated solida,b,c

Sulfuric acid

– 40.68 (1.44)

1.43 (0.16) 0.00 (0.00)

11.03 (0.66) 46.3 (2.89)

23.85 (1.21)

Sodium hydroxide

– 18.40 (0.16)

8.34 (0.15) 12.13 (0.40)

30.57 (0.56) 50.33 (1.84)

60.79 (2.75)

62.57 (2.57)

Hydrogen peroxide

– 25.59 (2.30)

0.90 (0.14) 10.00 (0.26)

17.21 (0.84) 34.53 (0.86)

49.82 (1.40)

7.78 (1.13)

Lignin

a b c d

Xylose

Glucan conversion (%)

Xylan conversion (%)

Glucose 0.00 (0.00)d

Composition percentages calculated from values on a dry-weight basis. Data are averages of three replicates. Numbers in parentheses represent standard deviations. Compositions of xylose and glucose in the hydrolysis supernatant are in upper rows while compositions of pretreated solids are in bottom rows. See text for explanation.

composition. The sulfuric acid and hydrogen peroxide pretreated samples had 2.2 times and 1.4 times the amount of lignin, respectively, compared to sodium hydroxide pretreated samples. Lignin is not attacked by the enzymes and therefore shields the cellulose during hydrolysis (Mansfield et al., 1999). Solubilization of xylan, on the other hand, seems to have a limited impact on cellulose digestibility. 3.7. Modeling Empirical quadratic models using time, temperature, and concentration as continuous variables and linear models relating a modified severity parameter to these variables were developed to predict xylan solubilization in sulfuric acid pretreatment and lignin reduction in sodium hydroxide pretreatment. These two treatment agents were chosen for modeling because they have been widely studied and seem to be the most promising pretreatments for use on cotton stalks. After eliminating the insignificant terms (p > 0.05) from the model based on the p-values from the Type III Sum of Squares ANOVA table (data not shown), the reduced empirical quadratic model used to quantify the percentage of xylan solubilized from the cotton stalks during sulfuric acid pretreatment was

The appropriate values for C, T, and t were plugged into the equations and the plots between fitted vs. observed values, for both percent xylan solubilization (Eq. (5)) and percent lignin reduction (Eq. (6)), had slopes of 0.97. Both models had high R2-values and slopes close to 1 thus indicating good agreement between the experimental data and the models. Linear models relating a modified severity parameter that combines the effects of time, temperature and concentration to the percentage solubilization of xylan by sulfuric acid pretreatment and to the reduction in lignin by sodium hydroxide pretreatment resulted in R2 values of 0.89 and 0.78, respectively. The n-values for sulfuric acid and sodium hydroxide pretreatments that provided the best model fits while keeping log (M0) positive were 0.849 and 3.90, respectively. The resulting equations were   T r  100 0:849 exp M 0 ðsulfuric acidÞ ¼ tC ð7Þ 14:75   T r  100 ð8Þ M 0 ðsodium hydroxideÞ ¼ tC 3:90 exp 14:75 The model equation for determination of xylan solubilization during sulfuric acid pretreatment was developed by plotting log (M0) vs. % xylan solubilization (Fig. 4).

120

þ 0:2644t  22:6728C þ 0:6347CT  11:0451C

2

ð5Þ

The square of the correlation coefficient (R2) for the xylan solubilization model was 0.964. The percent lignin reduction model for sodium hydroxide containing significant terms from the Type III Sum of Squares ANOVA table (data not shown) had an R2 of 0.924 and was given by % lignin reduction ¼ 1:3705 þ 0:0002T þ 0:5554t

y = 53.508x - 55.043 R2 = 0.8926

100 80 60 40 20 0 0.8

1.3

1.8

2.3

2.8

-20 log Mo

þ 49:6254C þ 0:0904Ct  15:9216C 2  0:0047t2

Xylan Solubilization (%)

Xylan solubilization ð%Þ ¼ 117:6194 þ 1:0798T

ð6Þ

Fig. 4. Percent xylan solubilization vs. log (modified severity parameter) for sulfuric acid pretreatment.

3010

R.A. Silverstein et al. / Bioresource Technology 98 (2007) 3000–3011

70

y = 8.6438x + 33.68 R2 = 0.7826

Lignin Reduction (%)

65 60 55 50 45 40 35 30 25 0

0.5

1

1.5

2 log Mo

2.5

3

3.5

4

Fig. 5. Percent lignin reduction vs. log (modified severity parameter) for sodium hydroxide pretreatment.

M0 was calculated using n = 0.849 and the model is represented by Eq. (9) % xylan solubilization ¼ 53:508  logðM 0 Þ  55:043

ð9Þ

The model equation for the reduction of lignin during sodium hydroxide pretreatment, using n = 3.90 to calculate M0, was obtained from Fig. 5 and is represented as % lignin reduction ¼ 8:6438  logðM 0 Þ þ 33:68

to water at high temperatures. Ozone pretreatment also did not perform as effectively as expected. Possible explanations include insufficient time, low ozone concentration, or uneven distribution of ozone throughout the sample. Compared with other pretreatments, sodium hydroxide pretreatment resulted in significantly (p < 0.05) higher cellulose conversion during the subsequent enzymatic hydrolysis. The empirical quadratic models successfully predicted percent xylan solubilization and percent lignin reduction and may be used in the development of better estimation tools. In addition, this study can serve as a step towards the optimization of pretreatment of cotton stalks. Nevertheless, different combinations of treatment factors, perhaps using higher temperatures or concentrations and application of higher pressures could be investigated. In addition, enzymatic hydrolysis using optimized pretreatment factors and ethanol fermentation need to be studied for bioethanol production since they could not be addressed in this study.

ð10Þ

The modified severity parameter model was validated by plotting the experimental values of percent xylan solubilization and percent lignin reduction against the model predicted values (Silverstein, 2004). The R2 from the plot of experimental vs. predicted % xylan solubilization was 0.88 with a slope of 0.95 indicating good predictive ability of the model. Predicted and experimental values for % lignin reduction during sodium hydroxide pretreatment resulted in an R2 of 0.72 and a slope of 0.99. Variation in predicted and experimental values may have likely been due to heterogeneity of cotton stalks and inability of the modified severity parameter to fully capture dependence of response variables on independent variables in the absence of variables such as stalk to cotton fiber ratio and solids loading. 4. Conclusions Sulfuric acid pretreatment substantially solubilized xylan in cotton stalks and temperature had the most significant effect on xylan solubilization. Data analysis indicated that there is a linearly increasing relationship between xylan solubilization and pretreatment severity. The most significant effect of sodium hydroxide pretreatment was on delignification with concentration of sodium hydroxide being the significant factor. Lignin reduction increased linearly with increase in pretreatment severity of sodium hydroxide. Hydrogen peroxide pretreatment resulted in lower lignin and xylan solubilization than expected. This was probably due to decomposition of hydrogen peroxide

References Agblevor, F.A., Evans, R.J., Johnson, K.D., 1994. Molecular-beam massspectrometric analysis of lignocellulosic materials. I. Herbaceous biomass. J. Anal. Appl. Pyrol. 30, 125–144. Agblevor, F.A., Batz, S., Trumbo, J., 2003. Composition and ethanol production potential of cotton gin residues. Appl. Biochem. Biotechnol. 105, 219–230. Azzam, A.M., 1989. Pretreatment of cane bagasse with alkaline hydrogen peroxide for enzymatic hydrolysis of cellulose and ethanol fermentation. J. Environ. Sci. Health B 24, 421–433. Badger, P.C., 2002. Ethanol from cellulose: a general review. In: Janick, J., Whipkey, A. (Eds.), Trends in New Crops and New Uses. ASHS Press, Alexandria, VA, pp. 17–21. Ben-Ghedalia, D., Shefet, G., Miron, J., 1980. Effect of ozone and ammonium hydroxide treatments on the composition and in vitro digestibility of cotton straw. J. Sci. Food Agric. 31, 1337–1342. Chang, V., Holtzapple, M., 2000. Fundamental factors affecting biomass enzymatic reactivity. Appl. Biochem. Biotechnol. 84–86, 5–37. Chum, H.L., Johnson, D.K., Black, S.K., Overend, R.P., 1990. Pretreatment-catalyst effects and the combined severity parameter. Appl. Biochem. Biotechnol. 24–25, 1–14. Ehrman, T., 1994. Method for determination of total solids in biomass. In: Laboratory Analytical Procedures No. 001. Golden, CO, National Renewable Energy Laboratory. Ehrman, T., 1996. Method for determination of acid-soluble lignin in biomass. In: Laboratory Analytical Procedures No. 004. Golden, CO, National Renewable Energy Laboratory. Gould, J.M., 1985. Studies on the mechanism of alkaline peroxide delignification of agricultural residues. Biotechnol. Bioeng. 27, 225–231. Han, J., Rowell, J., 1997. Chemical composition of fibers. In: Rowell, R., Young, R., Rowell, J. (Eds.), Paper Composites from Agro-Based Resources. CRC Lewis Publisher, New York, pp. 83–134. Hsu, T.A., 1996. Pretreatment of biomass. In: Wyman, C.E. (Ed.), Handbook on Bioethanol: Production and Utilization. Taylor & Francis, Washington, DC, pp. 179–195. Ingram, L.O., Doran, J., 1995. Conversion of cellulosic materials to ethanol. FEMS Microbiol. Rev. 16, 235–241. Kim, S.B., Um, B.H., Park, S.C., 2001. Effect of pretreatment of reagent and hydrogen peroxide on enzymatic hydrolysis of oak in percolation process. Appl. Biochem. Biotechnol. 91–93, 81–94.

R.A. Silverstein et al. / Bioresource Technology 98 (2007) 3000–3011 Mansfield, S.D., Mooney, C., Saddler, J.N., 1999. Substrate and enzyme characteristics that limit cellulose hydrolysis. Biotechnol. Prog. 15, 804–816. Martel, P., Gould, J.M., 1990. Cellulose stability and delignification after alkaline hydrogen-peroxide treatment of straw. J. Appl. Poly. Sci. 39, 707–714. McKendry, P., 2002. Energy production from biomass (part 1): overview of biomass. Bioresour. Technol. 83, 37–46. McMillan, J.D., 1994. Pretreatment of lignocellulosic biomass. In: Himmel, M.E., Baker, J.O., Overend, R.P. (Eds.), Enzymatic Conversion of Biomass for Fuels Production. American Chemical Society, Washington, DC, pp. 292–324. Milne, T.A., Chum, H.L., Agblevor, F.A., Johnson, D.K., 1992. Standardized analytical methods. Biomass Bioenergy 2, 341–366. Moiser, N., Wyman, C., Dale, B., Elander, R., Lee, Y.Y., Holtzapple, M., Ladisch, M., 2005. Features of promising technologies for pretreatment of lignocellulosic biomass. Bioresour. Technol. 96, 673–686. Neely, W.C., 1984. Factors affecting the pretreatment of biomass with gaseous ozone. Biotechnol. Bioeng. 26, 59–65. Overend, R.P., Chornet, E., 1987. Fractionation of lignocellulosics by steam-aqueous pretreatments. Philos. Trans. R. Soc. Lond. 321, 523–536. Quesada, J., Rubio, M., Gomez, D., 1999. Ozonation of lignin rich solid fractions from corn stalks. J. Wood Chem. Tech. 19, 115–137. Ruiz, R., Ehrman, T., 1996. Determination of carbohydrates in biomass by high performance liquid chromatography. In: Laboratory Analytical Procedures No. 002. Golden, CO, National Renewable Research Laboratory. Schell, D.J., Farmer, J., Newman, M., McMillan, J.D., 2003. Dilutesulfuric acid pretreatment of corn stover in pilot-scale reactor – investigation of yields, kinetics, and enzymatic digestibilities of solids. Appl. Biochem. Biotechnol. 105, 69–85. Sharma, R.R., Demirci, A., Beuchat, L.R., Fett, W.F., 2002. Inactivation of Escherichia coli O157:H7 on inoculated alfalfa seeds with ozonated water and heat treatment. J. Food Prot. 65, 447–451.

3011

Shoemaker, S., 2004. Advanced biocatalytic processing of heterogeneous lignocellulosic feedstocks to a platform chemical intermediate (lactic acid ester). Final report for award number DE-FC02-99CH11007. (accessed September, 2006). Silverstein, R., 2004. A comparison of chemical pretreatment methods for converting cotton stalks to ethanol. MS thesis, North Carolina State University. Available from: . Sun, Y., Cheng, J.J., 2002. Hydrolysis of lignocellulosic materials for ethanol production: a review. Bioresour. Technol. 83, 1–11. Tarkov, H., Feist, W.C., 1969. A mechanism for improving the digestibility of lignocellulosic materials with dilute alkali and liquid ammonia. Adv. Chem. Ser. 95 (1), 197–218. TBWEF. 2004. Texas Boll Weevil Eradication Foundation (TBWEF). (accessed 29.09.04). Templeton, D., Ehrman, T., 1994. Determination of acid-insoluble lignin in biomass. In: Laboratory Analytical Procedures No. 003. Golden, CO, National Renewable Energy Laboratory. Torget, R., Walter, P., Himmel, M., Grohmann, K., 1991. Dilute acid pretreatment of corn residues and short-rotation woody crops. Appl. Biochem. Biotechnol. 28–29, 75–86. USDA 2004. National Agricultural Statistics Service Crop Production. (accessed May, 2005). Varga, E., Scengyel, Z., Recaey, K., 2002. Chemical pretreatments of corn stover for enhancing enzymatic digestibility. Appl. Biochem. Biotechnol. 98–100, 73–87. Yang, B., Wyman, C.E., 2004. Effect of xylan and lignin removal by batch and flowthrough pretreatment on the enzymatic digestibility of corn stover cellulose. Biotechnol. Bioeng. 86, 88–95. Yang, B., Boussaid, A., Mansfield, S.D., Gregg, D.J., Saddler, J.N., 2002. Fast and efficient alkaline peroxide treatment to enhance the enzymatic digestibility of steam-exploded softwood substrates. Biotechnol. Bioeng. 77, 678–684.

A comparison of chemical pretreatment methods for ...

xylan reduction (95.23% for 2% acid, 90 min, 121 ╟C/15 psi) but the lowest cellulose to glucose ... research on the development of alternative energy sources.

331KB Sizes 4 Downloads 275 Views

Recommend Documents

A comparison of chemical pretreatment methods for ...
conversion of lignocellulosic biomass to ethanol is, how- ever, more challenging ..... treatments, subsequent calculations and analysis of data in this study were ...

a comparison of methods for determining the molecular ...
2009). The spatial and mass resolution of the cosmological simulation is matched exactly to the spatial and mass resolution of the simulation with the fixed ISM. Due to computational expense, the fully self- consistent cosmological simulation was not

A Comparison of Methods for Finding Steady- State ...
The existing commercial tools suggest a brute-force approach i.e. simulation ..... to sampled-data modeling for power electronic circuits , IEEE. Transactions on ...

A Comparison of Clustering Methods for Writer Identification and ...
a likely list of candidates. This list is ... (ICDAR 2005), IEEE Computer Society, 2005, pp. 1275-1279 ... lected from 250 Dutch subjects, predominantly stu- dents ...

A Comparison of Methods for Finding Steady- State ...
a converter is described as a linear time-invariant system. (LTI). ... resonant converters the procedure to find the solution is not as easy as it may .... New set of Ai.

A comparison of numerical methods for solving the ...
Jun 12, 2007 - solution. We may conclude that the FDS scheme is second-order accurate, but .... −a2 − cos bt + 2 arctan[γ(x, t)] + bsin bt + 2 arctan[γ(x, t)]. − ln.

COMPARISON METHODS FOR PROJECTIONS AND.pdf
There was a problem previewing this document. Retrying... Download. Connect more apps... Try one of the apps below to open or edit this item. COMPARISON ...

Intrinsic Methods for Comparison of Corpora - raslan 2013
Dec 6, 2013 - syntactic analysis. ... large and good-enough-quality textual content (e.g. news ... program code, SEO keywords, link spam, generated texts,. . . ).

Comparison of Stochastic Collocation Methods for ...
Oct 30, 2009 - ... of 58800 volumes with a double cell size in the direction tangential ...... Factorial sampling plans for preliminary computational experiments, ...

Comparison of Four Methods for Determining ...
limit the access of lysostaphin to the pentaglycine bridge in the ... third glycines to the pentaglycine cross bridge. .... Lysostaphin: enzymatic mode of action.

Comparison of Camera Motion Estimation Methods for ...
Items 1 - 8 - 2 Post-Doctoral Researcher, Construction Information Technology Group, Georgia Institute of. Technology ... field of civil engineering over the years.

Comparison of Training Methods for Deep Neural ... - Patrick GLAUNER
Attracted major IT companies including Google, Facebook, Microsoft and Baidu to make ..... Retrieved: April 22, 2015. The Analytics Store: Deep Learning.

A comparison of thinning methods in red pine
Abstract: Long-term replicated experiments that contrast thinning method (dominant thinning, .... ameter, using data from two long-term replicated studies in.

Evaluation of Prolonged Antithrombotic Pretreatment ...
charge. We performed a telephone in- terview at 30 days. For patients report- ing cardiac symptoms, at least 1 clinical and electrocardiographic examina-.

Randomized comparison of three surgical methods ...
Randomized comparison of three surgical methods used at the time of vaginal ... segment (level I) results from loss of the uterosacral-car- dinal ligament support.

comparison of signal processing methods: ica and ...
Dec 15, 2011 - and the Wavelet transform as a processing tool, along with the FFT for ... Techniques for signal processing data shown in this paper are the ICA ...

Implementation and Comparison of Solution Methods ...
tion methods such as structured policy iteration or SPUDD. [Boutilier et al. ... influence of various factors such as the structure and degree ..... 1:Unstructured). 0.1.

Comparison of Tabulation Methods Used by Two 1985 ...
The arguments to support the justification for adopting the AFA policy of dropping ... Movement of Contestants Into and Out of Award Categories. Using Different ...

Comparison of EPI Distortion Correction Methods in ...
between the two gradient echo images divided by the echo time difference using tools from Oxford ... remote from sources of susceptibility variation. Arrows are .... Proceedings of the SMRM 11th Annual Meeting (1992), 4515. 2. P. Jezzard and ...

Comparison of EPI Distortion Correction Methods in ...
We acquired four axial DWI datasets in each subject: the phase encoding direction was either anterior/posterior (A/P), or right/left (R/L), and the phase-phase encode blips had either positive or negative sign resulting in either compression or expan