ARTICLE pubs.acs.org/EF
Lime Pretreatment of Coastal Bermudagrass for Bioethanol Production Ziyu Wang and Jay J. Cheng* Department of Biological and Agricultural Engineering, North Carolina State University, Raleigh, North Carolina 27695-7625, United States ABSTRACT: Coastal bermudagrass (CBG) is regarded as a potential lignocellulosic feedstock for bioethanol production in the southeast United States. Lime pretreatment of CBG for enhanced reducing sugar recovery was investigated in this study, which examined a variety of temperatures (21121 °C) at a range of residence times with different lime loadings (0.020.20 g/g of dry biomass). During pretreatment, 1020% lignin was removed. After enzymatic hydrolysis with excessive cellulases and cellobiase, the best total reducing sugar yield for the lime-pretreated CBG was 78% of the theoretical maximum, which is over 2 times more than that from the untreated CBG. The recommended condition is 100 °C for 15 min with a lime loading of 0.1 g/g of dry biomass, under which 87% glucan and 68% xylan were converted to glucose and xylose, respectively. Fermentation tests of the hydrolyzates indicated that more than 99% glucose in the hydrolyzate was used by the yeast during the fermentation, with ethanol yields of 95% of the theoretical maximum for the hydrolyzate and 83% of the theoretical maximum for the raw biomass.
1. INTRODUCTION Ethanol, as a renewable transportation fuel, has the potential to replace gasoline for the reason that the use of fuel ethanol can alleviate the dependency upon fossil fuel, reduce net CO2 emissions, and slow down global climate change. The most widely used commercial conversion process for ethanol production in the United States is starch-based processing technology, in which corn is used as the feedstock. The continued increase in ethanol production using corn-based technology may not be sustainable because of limited agricultural land needed for food and feed production. However, enormous amounts of renewable lignocellulosic biomass are available for fuel ethanol production in addition to corn. Lignocellulose, the most abundant biomass, has been considered as a potentially inexpensive feedstock for ethanol production.1,2 It was projected that the equivalent of 40% of the U.S. gasoline consumption could be replaced by converting agricultural, forest, and municipal wastes alone to ethanol and the rest could be substituted by growing energy crops.3 The Energy Independence and Security Act of 2007 requires the use of 16 billion gallons of ethanol produced from cellulosic biomass annually in the U.S. by 2022.4 Therefore, there is a great interest to investigate cost-effective conversion processes for cellulosic ethanol production. The digestibility of lignocellulosic biomass during enzymatic hydrolysis is hindered by its structural features, such as cellulose crystallinity, lignin content, hemicellulose acetylation, and inaccessible surface area.5 Pretreatments are essential to enhance the susceptibility of lignocellulose to enzymatic hydrolysis by removing lignin and hemicellulose, reducing cellulose crystallinity, and increasing the porosity of the materials.6 Extensive research has been performed on various pretreatment methods, such as steam explosion, liquid hot water, ammonia, dilute acid, and alkaline pretreatments.714 The most commonly used alkalis are sodium hydroxide (NaOH), ammonia (NH3), and lime r 2011 American Chemical Society
[Ca(OH)2]. Fundamental studies indicated that alkalis remove lignin and acetate groups from hemicellulose, thus improving the reactivity of the remaining carbohydrates.5 In comparison to sodium hydroxide and ammonia, lime is cheaper, safer, and can be recovered by carbonating wash water with CO2.15 To make lime as efficient as other alkalis in enhancing the digestibility of lignocellulose, appropriate pretreatment conditions need to be employed. Studies have been carried out with lime pretreatment to optimize pretreatment conditions for various feedstocks: switchgrass, 100 °C for 2 h;13 wheat straw/bagasse, 85/120 °C for 3/1 h;16 corn stover, 120 °C for 4 h;17 and poplar wood/ newspaper, 150/140 °C for 6/3 h, with 14/7.1 atm oxygen.15 Adding air/oxygen to the reaction system can significantly improve the delignification of the biomass.18 Chang et al.15 performed oxidative lime pretreatment of poplar wood at 150 °C for 6 h with 78% removal of lignin and 71% improvement of the glucose yield from enzymatic hydrolysis. Lime (0.5 g of lime/g of raw biomass) was also used to pretreat corn stover in non-oxidative and oxidative conditions at 25, 35, 45, and 55 °C. The optimal condition was found to be 55 °C for 4 weeks with aeration.19 Coastal bermudagrass (CBG) (Cynodon dactylon L.) is widely grown in the southeast United States to remove nitrogen and phosphorus for pollution prevention and to feed animals.10 The existence of a cropping system makes CBG a potential lignocellulosic feedstock for bioethanol production. Although extensive research has been performed on lime pretreatment for various feedstocks, no study has investigated using lime to pretreat CBG for ethanol production. In this study, the effects of lime pretreatment on the enzymatic hydrolysis of CBG were investigated at different pretreatment temperatures, residence times, and lime Received: January 17, 2011 Revised: March 21, 2011 Published: March 21, 2011 1830
dx.doi.org/10.1021/ef2000932 | Energy Fuels 2011, 25, 1830–1836
Energy & Fuels loadings. The optimal conditions were generated on the basis of the efficiency of lime pretreatment on improved sugar production. The final ethanol yield from the optimal pretreatment conditions and enzyme dosing was detected to estimate potential ethanol production from CBG on a larger scale.
2. EXPERIMENTAL SECTION 2.1. Biomass Preparation. Air-dried CBG was supplied from the North Carolina State University Central Crops Research Station (Clayton, NC) in June, 2007. The biomass was ground in a Thomas Wiley Laboratory Mill (model 4) with a sieve diameter of 2 mm and then stored in sealed plastic bags at room temperature (21 °C) until use for pretreatment. 2.2. Lime Pretreatment. CBG samples (3 g per replicate) were mixed with lime (calcium hydroxide) in deionized (DI) water (solid/ liquid ratio of 1:10) in sealed serum bottles and pretreated in an autoclave (model 3021, Amsco) for pretreatments at 121 °C and a water bath (model 205, Fisher Scientific) for pretreatments up to 100 °C. After pretreatment, the biomass was recovered by filtration through a porcelain Buchner funnel and washed with 200 mL of DI water/0.1 g of lime used. The wet solid residues were stored in a sealed plastic bag at 4 °C for enzymatic hydrolysis. Two sets of experiments were conducted. The first set of experiments was performed at 21, 50, 80, 100, and 121 °C for various residence times, with a lime loading of 0.1 g/g of dry biomass. The second set examined pretreatments with different lime loadings (0.02, 0.05, 0.08, 0.15, and 0.20 g/g of dry biomass) at the above temperatures for the optimal residence times obtained in the first set. 2.3. Enzymatic Hydrolysis. Cellulases (NS50013 cellulase complex) produced by Trichoderma reesei and cellobiase (NS50010 βglucosidase) produced by Aspergillus niger were obtained from Novozymes North America, Inc. (Franklinton, NC). The cellulase and cellobiase activities were determined to be 76.44 filter paper units (FPU)/mL (expressed as micromoles of glucose produced per minute, with filter paper as a substrate) and 283.14 cellobiase units (CBU)/mL (expressed as micromoles of cellobiose that is converted to glucose per minute, with cellobiose as a substrate), respectively.20 All of the pretreated biomass was hydrolyzed with excessive enzyme dosage (cellulases, 40 FPU/g of dry biomass; cellobiase, 70 CBU/g of dry biomass) in 250 mL Erlenmeyer flasks in a controlled environmental incubator shaker (model 25, New Brunswick Scientific) at 55 °C and 150 rpm. For all of the hydrolysis experiments, 0.5 g (dry basis) of pretreated biomass was immersed in 0.05 M sodium citrate buffer to maintain a pH of 4.8 with the total liquid volume of 15 mL. Sodium azide (0.3%, w/v) was added to the hydrolysis mixture to inhibit microbial growth. The hydrolysis was carried out for 72 h, after which the hydrolyzate was centrifuged (model 5810R, Eppendorf) at 4 °C and 4000 rpm for 15 min, and the supernatant was stored at 20 °C for further analysis. Enzymatic hydrolysis of untreated biomass was conducted as a control. 2.4. Fermentation. Yeast (Saccharomyces cerevisiae, ATCC 24859) supplied from the Agricultural and Biological Engineering Department at Pennsylvania State University was aerobically grown at 30 °C in 100 mL of sterilized medium (consisting of 20 g of glucose, 8.5 g of yeast extract, 1.32 g of NH4Cl, 0.11 g of MgSO4, and 0.06 g of CaCl2 per liter of DI water) in a shaker incubator (model C76, New Brunswick Scientific) with 150 rpm for 24 h. The yeast cells were collected by centrifugation at 4000g at 4 °C for 10 min, washed 3 times with 0.1% peptone water, and then resuspended in 30 mL of peptone before use. A total of 1 mL of resuspended yeast sample was taken to measure the dry matter (%) of the inoculum, which was then used to determine the volume of yeast used to inoculate the hydrolyzate.
ARTICLE
After enzymatic hydrolysis, 5 mL of supernatant of the harvested hydrolyzate from each hydrolysis sample was added to 100 mL serum bottles for fermentation. The hydrolyzate was sterilized in an autoclave (model 3021, Amsco) at 121 °C/15 psi for 15 min, then adjusted to pH 7 by adding 2 N NaOH, and inoculated with S. cerevisiae at a cell concentration of 10 g of dry matter/L.21 All samples were then incubated (model 310, Instruments) at 30 °C for 72 h and analyzed for ethanol content after fermentation. The serum bottles were sealed during the fermentation without mixing. 2.5. Analytical Methods. The raw and pretreated biomass were analyzed for the contents of total solids, lignin, structural carbohydrates, and ash using Laboratory Analytical Procedures (LAP) established by National Renewable Energy Laboratory (NREL).2224 Total reducing sugar in the raw biomass and the enzymatic hydrolyzates was determined by the dinitrosalicylic acid (DNS) method using glucose as the standard.25 Monosaccharides (glucose, xylose, arabinose, and galactose) in the biomass (raw and pretreated) and hydrolyzates were measured with a high-performance liquid chromatography (HPLC) system (model 1200, Agilent). The HPLC system was equipped with a BioRad Aminex HPX-87P column (300 7.8 mm) tailored for analysis of hexoses and pentoses in lignocellulosic materials, a Bio-Rad MicroGuard column, a thermostatted autosampler, a quaternary pump, and a refractive index detector. The standards used were monomeric sugar at concentrations of 0.25, 2.5, 5.0, 7.5, and 10.0 g/L. The analytical column was operated at 80 °C with Milli-Q water (0.2 μm filtered) as the mobile phase at a flow rate of 0.6 mL/min. The samples were injected at 10 μL, and the acquisition time was 35 min. To allow for late-eluting compounds to come off the column, a post-run time of 25 min was included between injections. Ethanol in the fermentation liquor was determined with a different HPLC system (model 1200, Agilent) equipped with a Bio-Rad Aminex HPX-87H column (300 7.8 mm) used for analysis of carbohydrates in solution with alcohols, a Bio-Rad Micro-Guard column, and a refractive index detector. The analytical column was operated at 65 °C with 0.005 M H2SO4 (0.2 μm filtered) as the mobile phase at a flow rate of 0.7 mL/ min. The samples were injected at 10 μL, and the acquisition time was 25 min. Two ways for calculating ethanol yield in terms of the percentage of the theoretical maximum ethanol yield were reported. One was to determine the ethanol yield of the hydrolyzate. The other was to estimate the ethanol yield of the raw biomass. The following formulas were used to calculate the ethanol yields: ethanol yield of hydrolyzate ð%Þ
¼
grams of ethanol produced 100 grams of glucose in the hydrolyzate 0:511
ethanol yield of raw biomass ð%Þ
grams of ethanol produced=grams of raw biomass density of ethanol ðg=LÞ ¼ 100 theoretical ethanol yield ðL=g of raw biomassÞ To estimate a potential ethanol yield, a published value of 0.463 g of ethanol produced/g of (glucose þ xylose)26 was used to substitute the value of grams of ethanol produced per gram of glucose in the above formulas. 2.6. Statistical Analysis. Experimental data were statistically analyzed using the GLM procedure in SAS 8.02 software. Significant (p < 0.05) and non-significant differences between treatments were evaluated by Tukey adjustment for comparisons. All treatments were conducted in triplicates.
3. RESULTS AND DISCUSSION 3.1. Biomass Composition. In this study, we used the same batch of feedstock as that in the previous work of sodium hydroxide 1831
dx.doi.org/10.1021/ef2000932 |Energy Fuels 2011, 25, 1830–1836
Energy & Fuels
ARTICLE
pretreatment,14 and the composition of the CBG is shown in Table 1. Glucan was the major component, followed by xylan and acid-insoluble lignin. There were only a small amount of arabinan and galactan in the CBG. No mannan was detected in the biomass. Xylan was the main component of hemicellulose in the CBG. 3.2. High-Temperature Lime Pretreatment. The effects of sodium hydroxide pretreatment on enzymatic digestibility of CBG were investigated prior to the study of lime pretreatment.14 The results indicated that NaOH pretreatment at high temperature (121 °C) performed well on enhancing sugar production. For the lime pretreatment study, high-temperature (121 °C) pretreatments were explored first. A lime loading of 0.1 g/g of dry biomass was used according to the findings in a previous study on switchgrass.13 Comparisons between lime (0.1 g/g of dry biomass) pretreatment and NaOH (1%, equivalent to 0.1 g/g of dry biomass at the solid loading of 10%) pretreatment of CBG at 121 °C were made in the study. The results showed that Table 1. Chemical Composition of CBG14 component glucan
25.59
xylan
15.88
arabinan
1.95
galactan
1.46
acid-insoluble lignin acid-soluble lignin extractives ash othersa a
wt %, dry basis in biomass
15.37 3.96 4.17 6.6 25.02
Other components may include some organic compounds, minerals, waxes, fats, starches, resins, and gums.27,28
NaOH pretreatment was more effective in removing lignin than lime pretreatment, with an average difference of 55% lignin removal (Figure 1a). This is probably because sodium hydroxide is a stronger base than lime, which has a poor solubility (1.73 g/L at 20 °C) in water. Sodium hydroxide is able to give a higher pH than lime when the same amounts of the two chemicals are added to the same amount of water, thus leading to greater solubilization of lignin during NaOH pretreatment. There was no significant (p > 0.05) difference on lignin removal among residence times of 15, 30, and 60 min for lime pretreatment at 121 °C. Although the extent of delignification for lime pretreatment was much lower than that for NaOH pretreatment, there was no significant (p > 0.05) difference in total reducing sugar production between the two pretreatements at 121 °C (Figure 1b). The findings suggest that it is not necessary to remove up to 6080% lignin to enhance sugar production, which is in agreement with the results from previous studies,29,30 showing that the digestibility of the lignocellulosic biomass can be improved as long as the porosity is increased during the pretreatment to make the carbohydrates become more accessible to enzymes. As residence time increased from 15 to 60 min for lime pretreatment of CBG, total reducing sugar production declined by approximately 6%. This could be caused by a slight degradation of carbohydrates during pretreatments at longer residence times. Monomeric sugar analysis of hydrolyzates is shown in panels c and d of Figure 1. The glucose yield for lime (0.1 g/g of dry biomass) pretreated biomass was slightly less than that obtained for NaOH (1%) pretreatment at 121 °C. The potential maximum yield of glucose after enzymatic hydrolysis is dependent upon both the extent of disruption of crystalline cellulose and reduction of lignin content during the pretreatment.18 The crystallinity of cellulose is more easily disrupted by NaOH than lime during the pretreatment process probably because of the
Figure 1. (a) Total lignin reduction, (b) total reducing sugar production, (c) glucose yield, and (d) xylose yield from pretreated CBG with lime (0.1 g/g of dry biomass) and 1% (w/v) NaOH, respectively, at 121 °C. 1832
dx.doi.org/10.1021/ef2000932 |Energy Fuels 2011, 25, 1830–1836
Energy & Fuels
ARTICLE
Figure 2. Sugar production from pretreated CBG with lime (0.1 g/g of dry biomass) at (a) 21 °C, (b) 50 °C, (c) 80 °C, and (d) 100 °C.
higher pH in NaOH solution than in limewater at the same chemical concentration; thereby, more glucose can be released from crystalline cellulose for NaOH pretreatment. Because of the higher extent of delignification and decrystallization during NaOH pretreatment,12,14 hemicelluloses in the biomass have a greater chance to be degraded into oligomers and monomers (mainly xylose) that go into the pretreatment liquor. This may explain why lime pretreatment, to some extent, liberated more xylose than NaOH pretreatment after enzymatic hydrolysis, perhaps owing to less xylan loss during lime pretreatment. Because of the higher degradability of hemicellulose than cellulose, the xylose yield for both NaOH and lime pretreatment slightly decreased as the residence time extended from 15 to 60 min, while there was no significant (p > 0.05) difference in the glucose yields at the three residence times. This indicates that degradation of carbohydrates could occur as the pretreatment severity is raised. 3.3. Low-Temperature Lime Pretreatment. A previous study has shown that lime pretreatment works well at 50 °C compared to that at 121 °C.29 Although, in this study, the total reducing sugar yield for lime pretreatment at 121 °C could reach up to 74% of the theoretical maximum, which is comparable to 77% for NaOH pretreatment, it is favorable to study lime pretreatment at temperatures below 121 °C, with potential pretreatment cost savings because lime is more soluble in water at lower temperatures. The effect of lower temperatures ( 0.05) improvement of total reducing sugar production with the increase of lime loading (from 0.1 to 0.20 g/g of dry biomass) was observed (Figure 3a). As a result, there is no benefit from further increasing lime loading beyond 0.1 g/g of dry biomass. The total reducing sugar yield was reduced by 89% when decreasing lime loading from 0.1 to 0.08 g/g of dry biomass. Lime loadings of 0.02 and 0.05 g/g of dry biomass did not work at all in terms of sugar recovery. Lime is slightly soluble in water. The suspended lime particles have a very high total surface area, which means that, as the dissolved lime is used up during the pretreatment by rendering hydroxide ions to react with the biomass and forming lignincalciumlignin linkages,32 more lime particles will quickly dissolve into the solution. The dissolved divalent calcium ions in the solution will continue to cross-link lignin molecules, and the hydroxide ions will keep being consumed by the biomass. From Figure 3a, it indicates that any amount of lime more than 0.1 g was excessive for 1 g of dry biomass because the total reducing sugar production was not improved beyond the lime loading of 0.1 g/g of dry biomass. It was unexpectedly noted that, as lime 1834
dx.doi.org/10.1021/ef2000932 |Energy Fuels 2011, 25, 1830–1836
Energy & Fuels
ARTICLE
Table 3. Fermentation of Hydrolyzate from Enzymatic Hydrolysisa of Lime-Pretreatedb CBG ethanol yield
value
g/g of glucose
0.49
percentage of the theoretical yield for hydrolyzate (glucose) percentage of the theoretical yield for raw biomass (glucose)
96.7 82.9
g/g of (glucose þ xylose)c
0.463
percentage of the theoretical yield for hydrolyzate (glucose þ xylose)c90.6 percentage of the theoretical yield for raw biomass (glucose þ xylose)67.9 a
Optimal enzyme loadings (cellulase, 20 FPU/g; cellobiase, 10 CBU/g). b A total of 0.1 g/g of lime loading at 100 °C for 15 min. c From ref 26. Figure 4. Material balances from raw CBG to lime pretreatment at 50, 80, and 100 °C with a lime loading of 0.1 g/g of dry biomass for the optimal residence times (6 h at 50 °C, 3 h at 80 °C, and 15 min at 100 °C).
loading increased from 0.1 to 0.20 g/g of dry biomass, glucose and xylose yields both declined (panels b and c of Figure 3). This is probably related to the interactions of calcium ions with both lignin and carbohydrates in the biomass.29,32 It was reported that lignin exhibits a high affinity to calcium ions in solution32 and calcium ions could also congregate and precipitate lignin in a base medium,33 which explained why low lignin removal occurred during lime pretreatment of CBG. The results indicated that calcium ions, under excessive dosing conditions, probably formed more linkages with the biomass constituents, which consequently had negative impacts on the digestibility of cellulose and hemicellulose in the CBG. Although excessive lime loading negatively affected the release of glucose and xylose from the biomass because of extensive linkages formed among calcium ions, lignin, and carbohydrates, it did not have a negative impact on total reducing sugar production. This is probably because total reducing sugar includes not only glucose and xylose but also some oligosaccharides with reducing ends, such as cellobiose, which are normally easier to be formed during hydrolysis than the monosaccharides. 3.5. Material Balances. Mass balance was performed on the biomass pretreated under the optimal residence times at 50, 80, and 100 °C with a lime loading of 0.1 g/g of dry biomass. The total dry weight of the sample was measured after the pretreatment to determine total solid recovery. The contents of the pretreated biomass components (lignin, glucan, xylan, arabinan, galactan, ash, and others) were quantified and compared to those of the raw biomass. At all three combinations of pretreatment conditions, total solid recovery reached more than 80% (Figure 4), indicating a good overall preservation of the biomass after lime pretreatment. Further examination of lignin, glucan, and xylan shows that over 90% of glucan remained after the pretreatment but lignin and xylan were not preserved as well as glucan. Lime pretreatment resulted in a relatively low lignin reduction (
