- Email: [email protected]
Lime pretreatment to improve methane production of smooth cordgrass (Spartina alterniflora)
Chemical Engineering Journal 217 (2013) 337–344
Contents lists available at SciVerse ScienceDirect
Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej
Lime pretreatment to improve methane production of smooth cordgrass (Spartina alterniflora) Yue-gan Liang a, Zheng Zheng b,⇑, Xing-zhang Luo b, You-bin Si a, De-ju Cao a, Er Nie b, Beijiu Cheng c,⇑ a
School of Resource and Environment, Anhui Agricultural University, Hefei 230036, People’s Republic of China Department of Environmental Science and Engineering, Fudan University, Shanghai 200433, People’s Republic of China c Key Laboratory of Crop Biology of Anhui Province, Anhui Agricultural University, Hefei 230036, People’s Republic of China b
h i g h l i g h t s " Lime pretreatment increased 122–180% methane yield for pretreated samples vs. raw sample. " The highest methane yield (218.4 mL/g TS) was obtained under 0.12 g lime/g, 28 d and 45 °C. " Of all three factors, lime loading had the greatest influence on methane yield. " The condition of high lime loading, low temperature and short time was a preferential selection.
a r t i c l e
i n f o
Article history: Received 29 April 2012 Received in revised form 22 November 2012 Accepted 30 November 2012 Available online 8 December 2012 Keywords: Lime pretreatment Methane yield BMP Composition change Energy balance
a b s t r a c t Anaerobic conversion of smooth cordgrass (SC; Spartina alterniflora) to methane for energy production presents a viable option for effective management. The effect of lime pretreatment of SC on methane production was investigated in this study. The effect of three selected independent variables (lime loading (0.02–0.12 g Ca(OH)2/g SC), pretreatment time (7–28 days) and temperature (25–55 °C)) on material composition and methane production was explored. Lime pretreatment resulted in significant changes in SC composition: 5.7–60.5% hemicellulose and 10.2–36.2% lignin reductions, and 91–98.7% cellulose remaining in the solid residues after lime pretreatment. Lime pretreatment increased methane yield by between 122% and 180% and the methane production rate constant in the range of 56–212%. The highest methane yield was 218.4 mL/g total solids from pretreatment conditions of lime loading 0.12 g Ca(OH)2/g SC, for 28 days at 45 °C. These results were based on a 30-day biochemical methane potential assay. Lime loading had the greatest influence on methane yield of the three pretreatment factors. Considering the energy balance, the optimization of pretreatment experiment was carried out under the condition of 0.10 g Ca(OH)2/g SC dry mass, at 25 °C for 14 days. Based on this study, the pretreatment condition of high lime loading, low temperature and short pretreatment time was a preferential consideration for lime pretreatment. Ó 2012 Elsevier B.V. All rights reserved.
1. Introduction There have been significant researches on renewable sources of liquid fuels to replace fossil fuels to improve energy security and reduce greenhouse emissions. Lignocellulosic biomass, including crops specifically grown for energy, agricultural residues and the organic fraction of municipal solid waste, are all promising renewable resources, being widely available and convertible into various forms of fuel and chemical raw materials [1]. In China, since 2003, smooth cordgrass (SC; Spartina alterniflora) has been identified as ⇑ Corresponding authors. Tel./fax: +86 21 65643342 (Z. Zheng), tel.: +86 551 5786007; fax: +86 551 5786021 (B.J. Cheng). E-mail addresses: [email protected] (Y.-g. Liang), [email protected] (Z. Zheng), [email protected] (B.J. Cheng). 1385-8947/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.cej.2012.11.135
one of 16 invasive species by the China Environmental Protection Agency [2], because this plant has a rapid growth velocity and overwhelms the native ecosystems, resulting in clogging up rivers and even obstructing navigation in waterways [3]. The net primary production of SC varies from 110.9 to 601.8 g dry mass m2 year1, and by 2002 it was distributed over an area of at least 1120 km2 in China [3]. Anaerobic conversion of SC to methane as a clean fuel for energy production presents a viable option for effective management of SC. SC has been used as a potential and abundant biomass resource for producing biogas rich in CH4 by anaerobic digestion [2,4,5]. The lignocellulosic structure of SC is resistant to biodegradation and unitization for biofuels [1]. The crystallinity of cellulose, degree of polymerization, available surface area and shielding effect
338
Y.-g. Liang et al. / Chemical Engineering Journal 217 (2013) 337–344
of lignin all contribute to limiting the extent of enzymatic hydrolysis of biomass [6–8]. Pretreatment of lignocellulosic biomass is an essential and effective measure to disrupt these barriers [1]. Lime pretreatment is considered to be one of the most promising technologies for pretreatment of lignocellulosic biomass, because it is cheap, compatible with other oxidants, exhibits good carbohydrate preservation and is easy to recover [1,9,10]. Improvement of lignocellulosic biodegradation with lime pretreatment has been demonstrated on corn stovers [11], polar wood [10,12], sugarcane bagasse [13,14], and wheat straw [13], etc. Short-term or long-term lime pretreatments were recommended for high, medium or low lignin lignocelluloses contents [10,15]. Long-term lime pretreatment is preferred because it uses much a lower temperature and pressure, and pretreatment only needs a simple reactor, which may also be the biomass store [11,12]. The efficiency and effectiveness of pretreatment process not only depends on the pretreatment conditions, but also on the type of pretreated biomass [16]. However, relatively little has been published on the effect of lime pretreatment of the SC digestibility, and so in the present study long-term lime pretreatment was used to improve the methane production of SC. The effect of lime pretreatment on the SC was determined on the basis of methane yield by using the biochemical methane potential (BMP) assay. Chemical characterizations were performed to determine cellulose, hemicellulose and lignin contents for the raw and treated biomass, and the influence of different pretreatment conditions on methane yield was evaluated. A second-order polynomial model was used to describe the effect of the three selected independent variables (lime loading, pretreatment time and temperature) on methane yield, and energy and environment comparisons of different pretreatment conditions were also made. 2. Materials and methods 2.1. Smooth cordgrass (SC), lime and anaerobic seed cultures
Table 1 Uniform design experiment for the long-term lime pretreatment of SC. Pretreatment experiment
Lime loading (g Ca(OH)2/ g SC)
Time (d)
Temperature (°C)
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13
0.02 0.02 0.04 0.04 0.06 0.06 0.08 0.08 0.10 0.10 0.12 0.12 0
7 14 21 28 7 14 21 28 7 14 21 28 28
35 45 55 25 45 55 25 35 55 25 35 45 55
SC, smooth cordgrass.
(6 42), there were only 12 experiments to be done. To investigate the effect of temperature on the biodegradation of SC, a pretreatment experiment without adding lime was used as a control. So, total 13 pretreatment experiments were carried out and the experimental conditions are shown in Table 1. The operational method of lime pretreatment was chosen from standard protocols [18]. Thirty grams of SC (air-dried) and 350 mL distilled water were added into a 500 mL conical flask, and then the appropriate amount of calcium hydroxide was added to the flask and mixed thoroughly. The flask was covered with plastic film, which was held in place with a rubber band. Finally, the beaker was placed into an incubator at the set temperature and left for the time designated in Table 1. The flask contents were manually mixed twice a day to avoid stratification. When the pretreatment time had elapsed, the flask was removed from the incubator. The pH of the pretreatment liquor was immediately determined and total organic carbon (TOC) was analyzed by taking sample.
SC was collected from Dafeng County in Jiangsu Province, China, in September 2009, and was air-dried and chopped into 3–4 cm lengths using paper shears. The cellulose, hemicellulose and lignin contents were 32.9%, 34.2% and 10.9% and the C, H, N and O elemental contents were 43.9%, 6.3%, 0.5% and 49.3%, respectively. Total solids (TSs) of SC was 92.4% and volatile solids (VSs) was 92.4% of the TS, respectively. The SC sample was pretreated with lime (calcium hydroxide); the calcium hydroxide was of analytical grade, purchased from Nanjing chemical company of China. Anaerobic digestion cultures were seeded using digestion liquor from previous anaerobic digestions of SC [4]. The anaerobic cultures were filtered through 2 mm mesh size screen and concentrated by settling before being used as inoculums for the BMP assay. The TS content of seed cultures was 3.9% and VS content was 46% of the TS.
2.3. Sample preparation for BMP assay
2.2. Experiment design and pretreatment method
The ground samples of raw and pretreated SC were BMP assayed by using the method described by Owens and Chynoweth [20]. Two runs were to determine the methane potential of SC. For the assay, inoculums (50 mL) and SC (0.6 g SC) were placed into a 250 mL reagent bottle, and a stock solution with macro and micronutrients was added using the composition devised by Owens and Chynoweth [20]. Finally, a mixture containing NaHCO3 and KHCO3 (at a mass ratio of NaHCO3:KHCO3 = 2:3) to achieve an alkalinity of 3500 mg/L (as CaCO3) was added, and the bottle was filled up to 200 mL with CO2-free water. The reagent bottle was tightly sealed with a rubber stopper after the head space was flushed with nitrogen gas for 1 min. The reagent bottle was then placed into an incubator at 35 °C and a 30-day BMP assay
In this study, three factors (lime loading, pretreatment time, and pretreatment temperature) were chosen as critical variables. Design levels for lime loading, pretreatment temperature and time were selected according to the previous works [10,15,17]. Lime loading was between 0.02 and 0.12 (g Ca (OH)2/g dry SC), pretreatment time was between 7 and 28 days and the temperature was between 25 and 55 °C. The experimental range of pretreatment time and temperature was divided into four levels and the lime loading was divided into six levels. The pretreatment experiment was designed by using a uniform design method, proposed in 1980 by Fang et al. [17]. According to uniform design table U12
The samples of raw and pretreated SC were prepared according to the National Renewable Energy Laboratory standard procedure [19]. Pretreated SC residues were first washed with tap water for adjusting the pH and reducing the concentration of Ca ion before sample preparation, until the decanted water became colorless under adding 5–6 drops of 1% phenolphthalein solution as indicator, then dried at 45 °C for 48 h and cooled in a desiccator to room temperature, and finally the sample was ground and passed through a 100-mesh screen to prepare it for BMP assay. A raw SC sample was prepared by using the same drying and grinding procedure. 2.4. Biochemical methane potential (BMP) assay
Y.-g. Liang et al. / Chemical Engineering Journal 217 (2013) 337–344
was carried out according to the method proposed by Bilgili et al. [21]. Each bottle was manually mixed twice a day to avoid stratification. A control (with no additions) was given the same treatment to determine the background level of methane production from the inoculums. Specific methane yield was calculated according to the TS value of raw SC or pretreated solid residues used in BMP assay. Methane yields reported in this paper were the average value of two experimental values. 2.5. Analytical methods Daily biogas yield was recorded by the displacement of acidic saturated salt water and transformed to a value representing yield at standard temperature and pressure. The biogas composition was determined by gas chromatography (TCD detector, Porapak Q column, and helium carrier gas) using the analysis procedure suggested by Yang et al. [2]. TS and VS contents were determined according to standard methods [22]. The pH value was measured by using a pH meter (METER 6219), and the TOC value of the pretreated liquid was determined by using a TOC-5000A instrument made by Shimadzu, Japan. The detail method for TOC analysis was described by Yang et al. [2]. Neutral detergent fiber, acid detergent fiber and acid detergent lignin were analyzed according to the Van Soest method [23] with two replicates, and the value of cellulose, hemicellulose and lignin was calculated by using the method as described elsewhere [5]. C, H and N elemental contents were quantified by using an Element Analyzer (Heraeus Company, Germany), and the O content was calculated by difference. 2.6. Calculations and kinetic evaluation Removal efficiencies of TS, VS, cellulose, hemicellulose and lignin were calculated by Eq. (1), in which W0 and Wf were the weight of raw SC, pretreated solid residues, respectively. X0 and Xf were the value of raw SC and pretreated solid residues, including TS, VS, cellulose, hemicellulose and lignin contents, respectively. The dissolution part of pretreatment liquid, such as cellulose, hemicellulose and lignin contents was considered as the removed part. The recovery yield for cellulose was defined as (100% – removal efficiency). The value reported in this paper was the average value of two determination values.
Removal efficiency ¼
W 0 X0 W f Xf 100% W 0 X0
ð1Þ
A first-order kinetic model has been widely applied to anaerobic digestion to assess the performance of methane production for various biomasses under various different conditions [21,24,25]. In this study, the biodegradation of raw and pretreated SC was assumed to follow a first-order rate of decay. Thus, the cumulative methane yield was assumed to follow Eq. (2). A plot of ln (1B/ B0) against t yields the straight line given by Eq. (2), so the k value was determined from the slope of the line.
lnð1 Bt =B0 Þ ¼ kt
ð2Þ
Here Bt (mL/g) was the cumulative methane yield at digestion time t; B0 (mL/g) was the total accumulative methane yield assumed to equal the theoretical methane yield calculated by using the modified method of the literature [5]; k (d1) was the constant of first-order methane production rate, and t (d) was the digestion time. The theoretical methane yield was calculated according to the hemicellulose and cellulose contents by using Eq. (3). The theoretical methane yield of hemicellulose ((C5H8O4)n) and cellulose ((C6H10O5)n) was assumed to follow the Eqs. (4) and (5).
B0 ¼ B1 X 1 þ B2 X 2
ð3Þ
a b n a b Cn Ha Ob þ n H2 O ! þ CO2 4 2 2 8 4 n a b þ CH4 þ 2 8 4 Bt ¼
22:4 ð0:5n þ 0:125a 0:25bÞ 12n þ a þ 16b
339
ð4Þ
ð5Þ
Here B1 (mL/g) and B2 (mL/g) were the theoretical methane yield of cellulose and hemicellulose, respectively; X1 (%) and X2 (%) were the content of cellulose and hemicellulose for raw or pretreatment sample, respectively; Bt (mL/g) was the theoretical methane yield for the molecular formula of CnHaOb. 2.7. Statistical analysis and regression analysis A t-test was performed to check the data for any significant difference in terms of methane yield between pretreated and raw SC. The data were analyzed using SPSS 13.0 software (IBM-SPSS Inc.) with a confidence interval of 95%. A second-order polynomial model was used to describe the effect of the three selected independent variables on methane yields. Multiple regression analysis on the experimental data was performed using SPSS 13.0 software as above. This model may be expressed as:
B ¼ a0 þ
3 X X ai X i þ aij X i X j i¼1
ð6Þ
i6i6j63
Here B was the predicted value of methane yield; X was the independent variable, including lime loading, pretreatment time or temperature; and a was a regression coefficient estimated by the enter regression method. 3. Results and discussion 3.1. Effect of lime pretreatment on SC composition The pH and TOC values of the liquid fraction from lime pretreatment of SC are shown in Fig. 1a and b. The pH values of pretreatment liquor for P1–P4, P6, P8, and P13 were below 7, indicating that lime was completely consumed, mainly owing to lime disrupting ester bonds and neutralizing structural carboxylic acids formed through the deacetylation of hemicellulose during the lime pretreatment [9,11], and whereas the dissolution of hemicellulose occurred and carboxylic acids would generate from the cleavage of O-acetyl and uronic acid substitutions in hemicellulose part for P13 [9]. The TOC values of the pretreated liquor were very high, ranging from 12.2 to 34.7 g/L, and the composition of liquor was mainly carboxylic acids, phenols and furans and other solutes from lignin and hemicellulose [26]. A decrease in TOC value was found as the lime loading was increased. This is due to the conversion of a greater part of the dissolved lignin and hemicellulose into CO2 and H2O owing to the higher lime loading [26]. The high TOC content indicates that the pretreated liquid may be utilized for methane production. Methane production of 67.5 L/kg bagasse was reported from lime pretreatment liquor [14]. The composition changes in the solid residues after pretreatment are shown in Fig. 1c–e. TS and VS reductions fell in the ranges 8.5–30.2% and 3.9–29.6%, respectively. White mycelium was found on the surface of pretreated biomass for treatments P1, P3 and P4, and a high level of VS removal was also found in these experiments. Lime pretreatment resulted in 5.7–60.5% hemicellulose and 10.2–36.2% lignin reductions. The cellulose recovery yield reached 91–98.7%, except for experiment P1. This indicates that lime selectively removes hemicellulose or lignin. Similar results
340
Y.-g. Liang et al. / Chemical Engineering Journal 217 (2013) 337–344
Fig. 1. Changes in solid and liquor fraction after lime pretreatment.
were also reported with lime pretreatment of poplar wood [12]. The average 92% cellulose recovery yield was obtained from lime pretreatment of poplar wood [10], and Kim and Holtzapple [11] found that 93.7% of raw glucan was remained in the solid residues after non-oxidative lime pretreatment (0.5 g Ca(OH)2/g raw biomass, 55 °C, 16 weeks). Ninety-six percent cellulose was remained in the solid residues after alkaline wet oxidation (195 °C, 10 min, 6.5 g of Na2CO3) [26].
3.2. Effect of lime pretreatment on methane yield Methane yields from the BMP assay of pretreated and untreated SC are shown in Fig. 2. Compared with raw sample, methane yields from BMP assay increased by 122–180% for the pretreated samples. There was a significant statistical difference for levels of methane yield between pretreated SC and raw SC (one-sample t-test, t value = 9.081, p 6 0.001, df = 12). This indicates that the lime
341
Y.-g. Liang et al. / Chemical Engineering Journal 217 (2013) 337–344
Table 2 Significance of regression coefficient for methane yield from lime pretreatment of SC. Variables
Coefficient Standard error
348.4 Lime loading (X1) Pretreatment time (X2) 2.956 Pretreatment temperature (X3) 2.768 X1.X2 0.321 X1.X1 5009.6 X2.X2 0.128 X3.X3 0.034
Computed p-Value t-value
260.973 1.335 2.823 1.047 1.263 2.191 15.509 2.073 1886.862 2.655 .082 1.546 .018 1.883
0.230 0.335 0.071 0.084 0.038 0.173 0.109
R = 0.914, R2 = 0.835, Std. error of the estimate 13.7676. The bold value indicated a significance effect (p value < 0.1).
3.3. Regression analysis of methane production The regression equation of methane yield from pretreated and raw sample was calculated using the following equation: Fig. 2. Methane yield by anaerobic digestion of pretreated and untreated smooth cordgrass.
B ¼ 128:345 þ 348:4L 2:956t þ 2:768T 0:321Lt þ 5009:6L2 þ 0:128t 2 0:034T 2
pretreatment did enhance methane yield from SC. The highest yield occurred in the P12 experiment with 218.4 mL/g TS methane yield, using the following pretreatment conditions: lime loading 0.12 g Ca(OH)2/g SC, temperature 45 °C and time 28 days. Methane yields (206.6 mL/g TS) for the P10 experiment (0.10 g Ca(OH)2/g SC, 25 °C and 14 days) were slightly lower than that of P12. The P13 experiment (without adding lime, 55 °C and 28 days) also achieved a very high yield, reaching 202.5 mL/g TS methane yields. This is possibly due to a higher lignin removal with a greater level of polysaccharide retention in P13. Lignin was regarded as the major hindrance to complete hydrolysis of lignocelluloses, and lignin removal improved the passage of enzymes through lignin barriers to access the polysaccharides [27]. The removal of the hemicellulose backbone did not facilitate enzyme hydrolysis under high lime loading [28], and too much hemicellulose removal reduced the amount of potentially available substances to be consumed by the anaerobic microorganisms [29]. An example of a mass balance is shown in Fig. 3. For experiment P12, an initial 1.0 g air-dried SC contained 0.9236 g TS. After lime pretreatment, only 0.7880 g TS was left in the solid residues. A total of 172.1 mL CH4 was obtained from pretreated solid residues by a 30-day BMP assay. After pretreatment, the reduction of cellulose, hemicellulose and lignin contents reached 3.06%, 37.2% and 35.8%, respectively.
ð7Þ
B (mL/g) was the predicted methane production; L, t and T were the lime loading (g Ca(OH)2/g SC), pretreatment time (d) and temperature (°C), respectively. Table 2 shows the significance of regression coefficient for methane yield from lime pretreatment of smooth cordgrass. Significant effects were marked in bold (p value < 0.1). It could be seen that, at the 90% confidence level, lime loading, pretreatment time was not significant for methane yield. However, pretreatment temperature and quadratic lime loading as well as the interactions between lime loading/pretreatment time were significant. Quadratic lime loading was the most important factor affecting this response. The interaction between the lime loading and pretreatment time to methane yield was negative (Table 2 and Eq. (7)). The analysis of variance for the regression model (Eq. (7)) is given in Table 3. The determination of coefficient R2 = 0.835 indicates that 16.5% of total variation did not fit the model (Table 2). The statistical significance of the model equation was also confirmed by an F-test, where the value F = 3.109 exceeded the table value F (7, 6, 0.1) = 2.83. The Fisher F-test also indicates that the experimental results fitted the model well. Three-dimensional response plots are shown in Fig. 4. Increasing pretreatment time resulted in a significant decrease of methane yield; however, the methane yield increased when pretreatment time exceeded 21 days (Fig. 4a). A decrease of methane yield also occurred at a lower or higher pretreatment
Fig. 3. Example of mass balance and energy production for 12th pretreating experiment (0.12 g Ca(OH)2/g SC, 28 days at 45 °C).
342
Y.-g. Liang et al. / Chemical Engineering Journal 217 (2013) 337–344
Table 3 ANOVA for model.
Regression Residual Total
Sum of squares
df
Mean square
F
F(7,6,0.1)
p value
5744.076 1137.279 6881.354
7 6 13
820.582 189.546
3.109
2.83
0.047
2 weeks of the non-oxidative lime pretreatment [11]. The lower methane yield for medium-term pretreatment may be ascribed to a higher level of cellulose and hemicellulose removal. A lower cellulose and hemicellulose recovery was found at 2–6 weeks for non-oxidative and oxidative conditions at 55 °C [11]. Considering the three factors of pretreatment, lime loading had the largest influence on methane yield (Fig. 4 and Eq. (7)). 3.4. Lime pretreatment improving methane production rate constant
temperature. This is ascribed to less lignin removal at low temperatures and whereas a higher hemicellulose removal at high temperatures. Methane yield would increase with a rise in lime loading (Fig. 4b). This is ascribed to a higher level of lignin removal accompanied by sufficient deacetylation. The result was similar to the work by Chang et al. [13], who reported that the yield of reducing sugar increased when the lime loading was increased from 0.02 to 0.15 g Ca(OH)2/g dry biomass. From Fig. 4c, it was concluded that a short or long pretreatment time resulted in a higher methane yield, because most of the delignification happened in the first
The modeling results of methane production from raw and pretreated SC using a first-order kinetic model are shown in Table 4. All the plots showed a good linearity, with coefficients of determination greater than 0.84. The methane production rate constants (k) ranged from 0.064 to 0.128 d1. This was very close to the results of the BMP assays of switchgrass pretreated with microwaves (0.081–0.134 d1) [25] and Jatropha curcas L (0.069–0.138 d1) [24], and the k values in this study were higher than the results (0.01195–0.01571 d1) from digestion of landfilled solid wastes
Fig. 4. Response surface for the interactive effect on methane yield through BMP test (interactive effect of pretreatment time and temperature (a), lime loading and temperature (b), lime loading and pretreatment time (c)).
Y.-g. Liang et al. / Chemical Engineering Journal 217 (2013) 337–344 Table 4 Methane production rate constants and recovery rate of raw and pretreated SC. Run
K (d1)
R2
B0a (mL/g)
Methane recovery efficiency (%)
Raw P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13
0.041 0.070 0.064 0.066 0.064 0.074 0.070 0.076 0.092 0.112 0.128 0.119 0.108 0.066
0.844 0.974 0.936 0.968 0.944 0.965 0.920 0.938 0.937 0.987 0.949 0.978 0.990 0.966
281.4 291.4 297.0 282.8 293.0 286.3 305.9 251.7 251.6 247.9 256.0 223.5 266.5 323.3
45.1 64.1 56.5 56.5 57.3 64.7 59.0 72.7 76.9 77.2 83.4 86.0 81.4 62.6
R2, coefficients of determination; Methane recovery efficiency = the experimental methane yield 100%/the theoretical methane yield. a Theoretical methane yield.
[21], owing to a smaller particle size being used. Particle size reduction increases available specific surface and decreases the degree of polymerization [30]. These factors were all attributed to improve the hydrolysis rate of raw and pretreated SC. Compared with the k value of raw SC, the k value of pretreated SC increased between 56% and 212%. The highest k value happened with treatment P10, reaching 0.128 d1. The methane recovery efficiency was higher (56.5–86%) for pretreated SC when compared with raw SC (45.1%) (Table 4).
3.5. Economic and environmental comparison The method of pretreatment has been considered as a key step in the production of biofuels from lignocelluloses [1]. However, pretreatment is also one of the most energy intensive steps in the process [9]. Optimization of energy consumption was essential for the use of lime as a pretreatment to anaerobic digestion for the process to be economically feasible; therefore, a basic energy balance was preformed and the result is shown in Table S1. The samples from lime pretreatment produced a higher amount of methane compared to the untreated or raw sample. However, an additional energy with the value of 3.72–49.69 kJ/g TS for pretreatment biomass would be required, including the energy of heating pretreatment mixture and heat loss. Input energy was much higher than the increased energies provided by methane production enhancement from pretreated solid residues, yielding net energy balance of 0.42 to 46.81 kJ/g TS of biomass (Table S1). This implies that the increased energy production due to pretreatment was lower than the amount of input additional energy from pretreatment operation. So the reduction of heat loss energy and an additional energy production from anaerobic digestion of pretreatment liquid would be required. When the pretreatment liquid was subjected to anaerobic digestion, the energy of increased methane from pretreatment for P4, P7 and P10 could meet the energy requirement of pretreatment (Table S1). Considering the energy balance, the pretreatment condition of high lime loading, low temperature and short pretreatment time was a preferential selection, and in this study, the optimization of pretreatment experiment was P10 with pretreatment conditions of 0.10 g Ca(OH)2/g dry mass, at 25 °C for 14 days. Nevertheless, this was only a very rough analysis of energy. For a complete calculation, it should be necessary to take into account investment costs, maintenance costs and integration in the whole wastewater treatment process [31]. It should be noted that the aim of anaerobic digestion of SC not only was to obtain the energy, but also was to treat effectively the solid wastes of SC.
343
The major effect of lime pretreatment was the partly removal of lignin and hemicellulose from smooth cordgrass, thus improving the reactivity of the remaining polysaccharides [1]. However alkali extraction, including lime pretreatment could result in the solubilization, redistribution and condensation of lignin and the increscent in the crystalline state of the cellulose [32,33]. These effects would partly counteract the positive effects of lignin removal and cellulose swelling from pretreatment [34]. And simultaneously the residue of Ca2+ from pretreatment and the solubilized lignin components often had an inhibitory effect for methanogens [35], so the fermentation or anaerobic digestion of pretreated liquid from lime pretreatment required a conditioning treatment. P13 reduced the chemical costs compared to P1–P12 lime pretreatment since lime was not added, and simultaneously the pretreatment liquid of P13 also reduced the need for neutralization and conditioning chemicals. 4. Conclusions Lime pretreatment resulted in a significant change in SC compositions. The extent of change was a function of lime loading, pretreatment time and temperature. Lime pretreatment resulted in 5.7–60.5% hemicellulose and 10.2–36.2% lignin reductions, and good cellulose preservation with a 91–98.7% recovery yield. Lime pretreatment increased methane yield by 122–180% and the methane production rate constant by 56–212% for pretreated samples compared with raw sample. The highest methane yield with 218.4 mL/g TS was obtained from pretreatment conditions of lime loading 0.12 g Ca(OH)2/g SC, for 28 days at 45 °C. Lime loading had the greatest influence on methane yield of all the three pretreatment factors. Considering the energy balance, the optimization of pretreatment conditions was 0.10 g Ca(OH)2/g dry mass, at 25 °C for 14 days. Based on this study, the pretreatment condition of high lime loading, low temperature and short pretreatment time was a preferential consideration for lime pretreatment. Acknowledgements The work was partially supported by the China Ministry of Environmental Protection (2012ZX07102-004), Shanghai Science Foundation of China (11075039) and the Introducing and Stabilizing talents of Anhui Agricultural University (WD2012-2). Dr. Liang thanks Ms. Zhao Hong-fen for constant support and understanding. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.cej.2012.11.135. References [1] B. Yang, C.E. Wyman, Pretreatment: the key to unlocking low-cost cellulosic ethanol, Biofuels Bioprod. Bioref. 2 (2008) 26–40. [2] S.G. Yang, J.H. Li, Z. Zheng, Z. Meng, Characterization of Spartina alterniflora as feedstock for anaerobic digestion, Biomass Bioenergy 33 (2009) 597–602. [3] S.Q. An, B.H. Gu, C.F. Zhou, Z.S. Wang, Z.F. Deng, Y.B. Zhi, H.L. Li, L. Chen, D.H. Yu, Y.H. Liu, Spartina invasion in China: implications for invasive species management and future research, Weed Res. 47 (2007) 183–191. [4] G.Y. Chen, Z. Zheng, S.G. Yang, C.X. Fang, X.X. Zou, J.B. Zhang, Improve conversion of Spartina alterniflora into biogas by co-digestion with cow feces, Fuel Process. Technol. 91 (2010) 1416–1421. [5] Y.G. Liang, Z. Zheng, R.M. Hua, X.Z. Luo, A preliminary study of simultaneous lime treatment and dry digestion of smooth cordgrass for biogas production, Chem. Eng. J. 174 (2011) 175–181. [6] Y. Zhao, W.J. Lu, H.T. Wang, Supercritical hydrolysis of cellulose for oligosaccharide production in combined technology, Chem. Eng. J. 150 (2009) 411–417. [7] H.T. Tan, K.T. Lee, Understanding the impact of ionic liquid pretreatment on biomass and enzymatic hydrolysis, Chem. Eng. J. 183 (2012) 448–458.
344
Y.-g. Liang et al. / Chemical Engineering Journal 217 (2013) 337–344
[8] R. Haruna, M.K. Danquah, Enzymatic hydrolysis of microalgal biomass for bioethanol production, Chem. Eng. J. 168 (2011) 1079–1084. [9] N. Mosier, C. Wyman, B. Dale, R. Elander, Y.Y. Lee, M. Holtzapple, M. Ladisch, Features of promising technologies for pretreatment of lignocellulosic biomass, Bioresour. Technol. 96 (2005) 673–686. [10] R. Sierra, M.T. Holtzapple, C.B. Granda, Long-term lime pretreatment of poplar wood, AIChE J. 57 (2011) 1320–1328. [11] S. Kim, M.T. Holtzapple, Lime pretreatment and enzymatic hydrolysis of corn stover, Bioresour. Technol. 96 (2005) 1994–2006. [12] R. Sierra, L.A. Garcia, M.T. Holtzapple, Selectivity and delignification kinetics for oxidative and nonoxidative lime pretreatment of poplar wood, part III: long-term, Biotechnol. Progr. 26 (2010) 1685–1694. [13] V.S. Chang, M. Nagwani, M.T. Holtzapple, Lime pretreatment of crop residues bagasse and wheat straw, Appl. Biochem. Biotechnol. 74 (1998) 135–159. [14] S.C. Rabelo, H. Carrere, R.M. Filho, A.C. Costa, Production of bioethanol, methane and heat from sugarcane bagasse in a biorefinery concept, Bioresour. Technol. 102 (2011) 7887–7895. [15] H. Rughoonundun, C. Granda, R. Mohee, M.T. Holtzapple, Effect of thermochemical pretreatment on sewage sludge and its impact on carboxylic acids production, Waste Manage. 30 (2010) 1614–1621. [16] A.A. Modenbach, S.E. Nokes, The use of high-solids loadings in biomass pretreatment—a review, Biotechnol. Bioeng. 109 (2012) 1430–1442. [17] K.T. Fang, D.K.J. Lin, P. Winker, Y. Zhang, Uniform design: theory and application, Technometrics 42 (2000) 237–248. [18] R. Sierra, C.B. Granda, M.T. Holtzapple, Lime pretreatment, in: J.R. Mielenz (Ed.), Biofuels Methods and Protocols, Humana Press–Springer Protocols, New York, 2009, pp. 115–124. [19] B. Hames, R. Ruiz, C. Scarlata, A. Sluiter, J. Sluiter, D. Templeton, Preparation of samples for compositional analysis, Laboratory Analytical Procedure, National Renewable Energy Laboratory, 2008. [20] J.M. Owens, D.P. Chynoweth, Biochemical methane potential of municipal solid waste (MSW) components, Water Sci. Technol. 27 (1993) 1–14. [21] M.S. Bilgili, A. Demir, G. Varank, Evaluation and modeling of biochemical methane potential (BMP) of landfilled solid waste: a pilot scale study, Bioresour. Technol. 100 (2009) 4976–4980. [22] APHA, Standard methods for the examination of water and wastewater, 19th ed., Washington, DC, 1995.
[23] P.J. Van Soest, Use of detergents in the analysis of fibrous feeds. II. A rapid method for the determination of fiber and lignin, J. Assoc. Off. Anal. Chem. 46 (1963) 829–835. [24] V.N. Gunaseelan, Biomass estimates, characteristics, biochemical methane potential, kinetics and energy flow from Jatropha curcus on dry lands, Biomass Bioenergy 33 (2009) 589–596. [25] D. Jackowiak, J.C. Frigon, T. Ribeiro, A. Pauss, S. Guiot, Enhancing solubilisation and methane production kinetic of switchgrass by microwave pretreatment, Bioresour. Technol. 102 (2011) 3535–3540. [26] H.B. Klinke, B.K. Ahring, A.S. Schmidt, A.B. Thomsen, Characterization of degradation products from alkaline wet oxidation of wheat straw, Bioresour. Technol. 82 (2002) 15–26. [27] V.S. Chang, M.T. Holtzapple, Fundamental factors affecting biomass enzymatic reactivity, Appl. Biochem. Biotechnol. 84–86 (2000) 5–37. [28] K. Grohmann, D.J. Mitchell, M.E. Himmel, B.E. Dale, H.A. Schroeder, The role of ester groups in resistance of plant cell wall polysaccharides to enzymatic hydrolysis, Appl. Biochem. Biotechnol. 20–21 (1989) 45–61. [29] P. Alvira, E. Tomás-Pejó, M. Ballesteros, M.J. Negro, Pretreatment technologies for an efficient bioethanol production process based on enzymatic hydrolysis: a review, Bioresour. Technol. 101 (2010) 4851–4861. [30] L.M. Palmowski, J.A. Müller, Influence of the size reduction of organic waste on their anaerobic digestion, Water Sci. Technol. 41 (2000) 155–162. [31] C. Bougrier, J.P. Delgenès, H. Carrère, Impacts of thermal pre-treatments on the semi-continuous anaerobic digestion of waste activated sludge, Biochem. Eng. J. 34 (2007) 20–27. [32] A.T.W.M. Hendriks, G. Zeeman, Pretreatments to enhance the digestibility of lignocellulosic biomass, Bioresour. Technol. 100 (2009) 10–18. [33] Y.G. Liang, Z. Zheng, X.Z. Luo, F.H. Guo, L.M. Wang, J.B. Zhang, Effect of mesophilic and thermophilic conditions on changes of physicochemical characteristics of smooth cordgrass via dry digestion process, Chem. Eng. J. 168 (2011) 544–552. [34] D. Gregg, J.N. Saddler, A techno-economic assessment of the pretreatment and fractionation steps of a biomass-to-ethanol process, Appl. Biochem. Biotechnol. 57–58 (1996) 711–727. [35] Y. Chen, J.J. Cheng, K.S. Creamer, Inhibition of anaerobic digestion process: a review, Bioresour. Technol. 99 (2008) 4044–4064.
Contents lists available at SciVerse ScienceDirect
Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej
Lime pretreatment to improve methane production of smooth cordgrass (Spartina alterniflora) Yue-gan Liang a, Zheng Zheng b,⇑, Xing-zhang Luo b, You-bin Si a, De-ju Cao a, Er Nie b, Beijiu Cheng c,⇑ a
School of Resource and Environment, Anhui Agricultural University, Hefei 230036, People’s Republic of China Department of Environmental Science and Engineering, Fudan University, Shanghai 200433, People’s Republic of China c Key Laboratory of Crop Biology of Anhui Province, Anhui Agricultural University, Hefei 230036, People’s Republic of China b
h i g h l i g h t s " Lime pretreatment increased 122–180% methane yield for pretreated samples vs. raw sample. " The highest methane yield (218.4 mL/g TS) was obtained under 0.12 g lime/g, 28 d and 45 °C. " Of all three factors, lime loading had the greatest influence on methane yield. " The condition of high lime loading, low temperature and short time was a preferential selection.
a r t i c l e
i n f o
Article history: Received 29 April 2012 Received in revised form 22 November 2012 Accepted 30 November 2012 Available online 8 December 2012 Keywords: Lime pretreatment Methane yield BMP Composition change Energy balance
a b s t r a c t Anaerobic conversion of smooth cordgrass (SC; Spartina alterniflora) to methane for energy production presents a viable option for effective management. The effect of lime pretreatment of SC on methane production was investigated in this study. The effect of three selected independent variables (lime loading (0.02–0.12 g Ca(OH)2/g SC), pretreatment time (7–28 days) and temperature (25–55 °C)) on material composition and methane production was explored. Lime pretreatment resulted in significant changes in SC composition: 5.7–60.5% hemicellulose and 10.2–36.2% lignin reductions, and 91–98.7% cellulose remaining in the solid residues after lime pretreatment. Lime pretreatment increased methane yield by between 122% and 180% and the methane production rate constant in the range of 56–212%. The highest methane yield was 218.4 mL/g total solids from pretreatment conditions of lime loading 0.12 g Ca(OH)2/g SC, for 28 days at 45 °C. These results were based on a 30-day biochemical methane potential assay. Lime loading had the greatest influence on methane yield of the three pretreatment factors. Considering the energy balance, the optimization of pretreatment experiment was carried out under the condition of 0.10 g Ca(OH)2/g SC dry mass, at 25 °C for 14 days. Based on this study, the pretreatment condition of high lime loading, low temperature and short pretreatment time was a preferential consideration for lime pretreatment. Ó 2012 Elsevier B.V. All rights reserved.
1. Introduction There have been significant researches on renewable sources of liquid fuels to replace fossil fuels to improve energy security and reduce greenhouse emissions. Lignocellulosic biomass, including crops specifically grown for energy, agricultural residues and the organic fraction of municipal solid waste, are all promising renewable resources, being widely available and convertible into various forms of fuel and chemical raw materials [1]. In China, since 2003, smooth cordgrass (SC; Spartina alterniflora) has been identified as ⇑ Corresponding authors. Tel./fax: +86 21 65643342 (Z. Zheng), tel.: +86 551 5786007; fax: +86 551 5786021 (B.J. Cheng). E-mail addresses: [email protected] (Y.-g. Liang), [email protected] (Z. Zheng), [email protected] (B.J. Cheng). 1385-8947/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.cej.2012.11.135
one of 16 invasive species by the China Environmental Protection Agency [2], because this plant has a rapid growth velocity and overwhelms the native ecosystems, resulting in clogging up rivers and even obstructing navigation in waterways [3]. The net primary production of SC varies from 110.9 to 601.8 g dry mass m2 year1, and by 2002 it was distributed over an area of at least 1120 km2 in China [3]. Anaerobic conversion of SC to methane as a clean fuel for energy production presents a viable option for effective management of SC. SC has been used as a potential and abundant biomass resource for producing biogas rich in CH4 by anaerobic digestion [2,4,5]. The lignocellulosic structure of SC is resistant to biodegradation and unitization for biofuels [1]. The crystallinity of cellulose, degree of polymerization, available surface area and shielding effect
338
Y.-g. Liang et al. / Chemical Engineering Journal 217 (2013) 337–344
of lignin all contribute to limiting the extent of enzymatic hydrolysis of biomass [6–8]. Pretreatment of lignocellulosic biomass is an essential and effective measure to disrupt these barriers [1]. Lime pretreatment is considered to be one of the most promising technologies for pretreatment of lignocellulosic biomass, because it is cheap, compatible with other oxidants, exhibits good carbohydrate preservation and is easy to recover [1,9,10]. Improvement of lignocellulosic biodegradation with lime pretreatment has been demonstrated on corn stovers [11], polar wood [10,12], sugarcane bagasse [13,14], and wheat straw [13], etc. Short-term or long-term lime pretreatments were recommended for high, medium or low lignin lignocelluloses contents [10,15]. Long-term lime pretreatment is preferred because it uses much a lower temperature and pressure, and pretreatment only needs a simple reactor, which may also be the biomass store [11,12]. The efficiency and effectiveness of pretreatment process not only depends on the pretreatment conditions, but also on the type of pretreated biomass [16]. However, relatively little has been published on the effect of lime pretreatment of the SC digestibility, and so in the present study long-term lime pretreatment was used to improve the methane production of SC. The effect of lime pretreatment on the SC was determined on the basis of methane yield by using the biochemical methane potential (BMP) assay. Chemical characterizations were performed to determine cellulose, hemicellulose and lignin contents for the raw and treated biomass, and the influence of different pretreatment conditions on methane yield was evaluated. A second-order polynomial model was used to describe the effect of the three selected independent variables (lime loading, pretreatment time and temperature) on methane yield, and energy and environment comparisons of different pretreatment conditions were also made. 2. Materials and methods 2.1. Smooth cordgrass (SC), lime and anaerobic seed cultures
Table 1 Uniform design experiment for the long-term lime pretreatment of SC. Pretreatment experiment
Lime loading (g Ca(OH)2/ g SC)
Time (d)
Temperature (°C)
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13
0.02 0.02 0.04 0.04 0.06 0.06 0.08 0.08 0.10 0.10 0.12 0.12 0
7 14 21 28 7 14 21 28 7 14 21 28 28
35 45 55 25 45 55 25 35 55 25 35 45 55
SC, smooth cordgrass.
(6 42), there were only 12 experiments to be done. To investigate the effect of temperature on the biodegradation of SC, a pretreatment experiment without adding lime was used as a control. So, total 13 pretreatment experiments were carried out and the experimental conditions are shown in Table 1. The operational method of lime pretreatment was chosen from standard protocols [18]. Thirty grams of SC (air-dried) and 350 mL distilled water were added into a 500 mL conical flask, and then the appropriate amount of calcium hydroxide was added to the flask and mixed thoroughly. The flask was covered with plastic film, which was held in place with a rubber band. Finally, the beaker was placed into an incubator at the set temperature and left for the time designated in Table 1. The flask contents were manually mixed twice a day to avoid stratification. When the pretreatment time had elapsed, the flask was removed from the incubator. The pH of the pretreatment liquor was immediately determined and total organic carbon (TOC) was analyzed by taking sample.
SC was collected from Dafeng County in Jiangsu Province, China, in September 2009, and was air-dried and chopped into 3–4 cm lengths using paper shears. The cellulose, hemicellulose and lignin contents were 32.9%, 34.2% and 10.9% and the C, H, N and O elemental contents were 43.9%, 6.3%, 0.5% and 49.3%, respectively. Total solids (TSs) of SC was 92.4% and volatile solids (VSs) was 92.4% of the TS, respectively. The SC sample was pretreated with lime (calcium hydroxide); the calcium hydroxide was of analytical grade, purchased from Nanjing chemical company of China. Anaerobic digestion cultures were seeded using digestion liquor from previous anaerobic digestions of SC [4]. The anaerobic cultures were filtered through 2 mm mesh size screen and concentrated by settling before being used as inoculums for the BMP assay. The TS content of seed cultures was 3.9% and VS content was 46% of the TS.
2.3. Sample preparation for BMP assay
2.2. Experiment design and pretreatment method
The ground samples of raw and pretreated SC were BMP assayed by using the method described by Owens and Chynoweth [20]. Two runs were to determine the methane potential of SC. For the assay, inoculums (50 mL) and SC (0.6 g SC) were placed into a 250 mL reagent bottle, and a stock solution with macro and micronutrients was added using the composition devised by Owens and Chynoweth [20]. Finally, a mixture containing NaHCO3 and KHCO3 (at a mass ratio of NaHCO3:KHCO3 = 2:3) to achieve an alkalinity of 3500 mg/L (as CaCO3) was added, and the bottle was filled up to 200 mL with CO2-free water. The reagent bottle was tightly sealed with a rubber stopper after the head space was flushed with nitrogen gas for 1 min. The reagent bottle was then placed into an incubator at 35 °C and a 30-day BMP assay
In this study, three factors (lime loading, pretreatment time, and pretreatment temperature) were chosen as critical variables. Design levels for lime loading, pretreatment temperature and time were selected according to the previous works [10,15,17]. Lime loading was between 0.02 and 0.12 (g Ca (OH)2/g dry SC), pretreatment time was between 7 and 28 days and the temperature was between 25 and 55 °C. The experimental range of pretreatment time and temperature was divided into four levels and the lime loading was divided into six levels. The pretreatment experiment was designed by using a uniform design method, proposed in 1980 by Fang et al. [17]. According to uniform design table U12
The samples of raw and pretreated SC were prepared according to the National Renewable Energy Laboratory standard procedure [19]. Pretreated SC residues were first washed with tap water for adjusting the pH and reducing the concentration of Ca ion before sample preparation, until the decanted water became colorless under adding 5–6 drops of 1% phenolphthalein solution as indicator, then dried at 45 °C for 48 h and cooled in a desiccator to room temperature, and finally the sample was ground and passed through a 100-mesh screen to prepare it for BMP assay. A raw SC sample was prepared by using the same drying and grinding procedure. 2.4. Biochemical methane potential (BMP) assay
Y.-g. Liang et al. / Chemical Engineering Journal 217 (2013) 337–344
was carried out according to the method proposed by Bilgili et al. [21]. Each bottle was manually mixed twice a day to avoid stratification. A control (with no additions) was given the same treatment to determine the background level of methane production from the inoculums. Specific methane yield was calculated according to the TS value of raw SC or pretreated solid residues used in BMP assay. Methane yields reported in this paper were the average value of two experimental values. 2.5. Analytical methods Daily biogas yield was recorded by the displacement of acidic saturated salt water and transformed to a value representing yield at standard temperature and pressure. The biogas composition was determined by gas chromatography (TCD detector, Porapak Q column, and helium carrier gas) using the analysis procedure suggested by Yang et al. [2]. TS and VS contents were determined according to standard methods [22]. The pH value was measured by using a pH meter (METER 6219), and the TOC value of the pretreated liquid was determined by using a TOC-5000A instrument made by Shimadzu, Japan. The detail method for TOC analysis was described by Yang et al. [2]. Neutral detergent fiber, acid detergent fiber and acid detergent lignin were analyzed according to the Van Soest method [23] with two replicates, and the value of cellulose, hemicellulose and lignin was calculated by using the method as described elsewhere [5]. C, H and N elemental contents were quantified by using an Element Analyzer (Heraeus Company, Germany), and the O content was calculated by difference. 2.6. Calculations and kinetic evaluation Removal efficiencies of TS, VS, cellulose, hemicellulose and lignin were calculated by Eq. (1), in which W0 and Wf were the weight of raw SC, pretreated solid residues, respectively. X0 and Xf were the value of raw SC and pretreated solid residues, including TS, VS, cellulose, hemicellulose and lignin contents, respectively. The dissolution part of pretreatment liquid, such as cellulose, hemicellulose and lignin contents was considered as the removed part. The recovery yield for cellulose was defined as (100% – removal efficiency). The value reported in this paper was the average value of two determination values.
Removal efficiency ¼
W 0 X0 W f Xf 100% W 0 X0
ð1Þ
A first-order kinetic model has been widely applied to anaerobic digestion to assess the performance of methane production for various biomasses under various different conditions [21,24,25]. In this study, the biodegradation of raw and pretreated SC was assumed to follow a first-order rate of decay. Thus, the cumulative methane yield was assumed to follow Eq. (2). A plot of ln (1B/ B0) against t yields the straight line given by Eq. (2), so the k value was determined from the slope of the line.
lnð1 Bt =B0 Þ ¼ kt
ð2Þ
Here Bt (mL/g) was the cumulative methane yield at digestion time t; B0 (mL/g) was the total accumulative methane yield assumed to equal the theoretical methane yield calculated by using the modified method of the literature [5]; k (d1) was the constant of first-order methane production rate, and t (d) was the digestion time. The theoretical methane yield was calculated according to the hemicellulose and cellulose contents by using Eq. (3). The theoretical methane yield of hemicellulose ((C5H8O4)n) and cellulose ((C6H10O5)n) was assumed to follow the Eqs. (4) and (5).
B0 ¼ B1 X 1 þ B2 X 2
ð3Þ
a b n a b Cn Ha Ob þ n H2 O ! þ CO2 4 2 2 8 4 n a b þ CH4 þ 2 8 4 Bt ¼
22:4 ð0:5n þ 0:125a 0:25bÞ 12n þ a þ 16b
339
ð4Þ
ð5Þ
Here B1 (mL/g) and B2 (mL/g) were the theoretical methane yield of cellulose and hemicellulose, respectively; X1 (%) and X2 (%) were the content of cellulose and hemicellulose for raw or pretreatment sample, respectively; Bt (mL/g) was the theoretical methane yield for the molecular formula of CnHaOb. 2.7. Statistical analysis and regression analysis A t-test was performed to check the data for any significant difference in terms of methane yield between pretreated and raw SC. The data were analyzed using SPSS 13.0 software (IBM-SPSS Inc.) with a confidence interval of 95%. A second-order polynomial model was used to describe the effect of the three selected independent variables on methane yields. Multiple regression analysis on the experimental data was performed using SPSS 13.0 software as above. This model may be expressed as:
B ¼ a0 þ
3 X X ai X i þ aij X i X j i¼1
ð6Þ
i6i6j63
Here B was the predicted value of methane yield; X was the independent variable, including lime loading, pretreatment time or temperature; and a was a regression coefficient estimated by the enter regression method. 3. Results and discussion 3.1. Effect of lime pretreatment on SC composition The pH and TOC values of the liquid fraction from lime pretreatment of SC are shown in Fig. 1a and b. The pH values of pretreatment liquor for P1–P4, P6, P8, and P13 were below 7, indicating that lime was completely consumed, mainly owing to lime disrupting ester bonds and neutralizing structural carboxylic acids formed through the deacetylation of hemicellulose during the lime pretreatment [9,11], and whereas the dissolution of hemicellulose occurred and carboxylic acids would generate from the cleavage of O-acetyl and uronic acid substitutions in hemicellulose part for P13 [9]. The TOC values of the pretreated liquor were very high, ranging from 12.2 to 34.7 g/L, and the composition of liquor was mainly carboxylic acids, phenols and furans and other solutes from lignin and hemicellulose [26]. A decrease in TOC value was found as the lime loading was increased. This is due to the conversion of a greater part of the dissolved lignin and hemicellulose into CO2 and H2O owing to the higher lime loading [26]. The high TOC content indicates that the pretreated liquid may be utilized for methane production. Methane production of 67.5 L/kg bagasse was reported from lime pretreatment liquor [14]. The composition changes in the solid residues after pretreatment are shown in Fig. 1c–e. TS and VS reductions fell in the ranges 8.5–30.2% and 3.9–29.6%, respectively. White mycelium was found on the surface of pretreated biomass for treatments P1, P3 and P4, and a high level of VS removal was also found in these experiments. Lime pretreatment resulted in 5.7–60.5% hemicellulose and 10.2–36.2% lignin reductions. The cellulose recovery yield reached 91–98.7%, except for experiment P1. This indicates that lime selectively removes hemicellulose or lignin. Similar results
340
Y.-g. Liang et al. / Chemical Engineering Journal 217 (2013) 337–344
Fig. 1. Changes in solid and liquor fraction after lime pretreatment.
were also reported with lime pretreatment of poplar wood [12]. The average 92% cellulose recovery yield was obtained from lime pretreatment of poplar wood [10], and Kim and Holtzapple [11] found that 93.7% of raw glucan was remained in the solid residues after non-oxidative lime pretreatment (0.5 g Ca(OH)2/g raw biomass, 55 °C, 16 weeks). Ninety-six percent cellulose was remained in the solid residues after alkaline wet oxidation (195 °C, 10 min, 6.5 g of Na2CO3) [26].
3.2. Effect of lime pretreatment on methane yield Methane yields from the BMP assay of pretreated and untreated SC are shown in Fig. 2. Compared with raw sample, methane yields from BMP assay increased by 122–180% for the pretreated samples. There was a significant statistical difference for levels of methane yield between pretreated SC and raw SC (one-sample t-test, t value = 9.081, p 6 0.001, df = 12). This indicates that the lime
341
Y.-g. Liang et al. / Chemical Engineering Journal 217 (2013) 337–344
Table 2 Significance of regression coefficient for methane yield from lime pretreatment of SC. Variables
Coefficient Standard error
348.4 Lime loading (X1) Pretreatment time (X2) 2.956 Pretreatment temperature (X3) 2.768 X1.X2 0.321 X1.X1 5009.6 X2.X2 0.128 X3.X3 0.034
Computed p-Value t-value
260.973 1.335 2.823 1.047 1.263 2.191 15.509 2.073 1886.862 2.655 .082 1.546 .018 1.883
0.230 0.335 0.071 0.084 0.038 0.173 0.109
R = 0.914, R2 = 0.835, Std. error of the estimate 13.7676. The bold value indicated a significance effect (p value < 0.1).
3.3. Regression analysis of methane production The regression equation of methane yield from pretreated and raw sample was calculated using the following equation: Fig. 2. Methane yield by anaerobic digestion of pretreated and untreated smooth cordgrass.
B ¼ 128:345 þ 348:4L 2:956t þ 2:768T 0:321Lt þ 5009:6L2 þ 0:128t 2 0:034T 2
pretreatment did enhance methane yield from SC. The highest yield occurred in the P12 experiment with 218.4 mL/g TS methane yield, using the following pretreatment conditions: lime loading 0.12 g Ca(OH)2/g SC, temperature 45 °C and time 28 days. Methane yields (206.6 mL/g TS) for the P10 experiment (0.10 g Ca(OH)2/g SC, 25 °C and 14 days) were slightly lower than that of P12. The P13 experiment (without adding lime, 55 °C and 28 days) also achieved a very high yield, reaching 202.5 mL/g TS methane yields. This is possibly due to a higher lignin removal with a greater level of polysaccharide retention in P13. Lignin was regarded as the major hindrance to complete hydrolysis of lignocelluloses, and lignin removal improved the passage of enzymes through lignin barriers to access the polysaccharides [27]. The removal of the hemicellulose backbone did not facilitate enzyme hydrolysis under high lime loading [28], and too much hemicellulose removal reduced the amount of potentially available substances to be consumed by the anaerobic microorganisms [29]. An example of a mass balance is shown in Fig. 3. For experiment P12, an initial 1.0 g air-dried SC contained 0.9236 g TS. After lime pretreatment, only 0.7880 g TS was left in the solid residues. A total of 172.1 mL CH4 was obtained from pretreated solid residues by a 30-day BMP assay. After pretreatment, the reduction of cellulose, hemicellulose and lignin contents reached 3.06%, 37.2% and 35.8%, respectively.
ð7Þ
B (mL/g) was the predicted methane production; L, t and T were the lime loading (g Ca(OH)2/g SC), pretreatment time (d) and temperature (°C), respectively. Table 2 shows the significance of regression coefficient for methane yield from lime pretreatment of smooth cordgrass. Significant effects were marked in bold (p value < 0.1). It could be seen that, at the 90% confidence level, lime loading, pretreatment time was not significant for methane yield. However, pretreatment temperature and quadratic lime loading as well as the interactions between lime loading/pretreatment time were significant. Quadratic lime loading was the most important factor affecting this response. The interaction between the lime loading and pretreatment time to methane yield was negative (Table 2 and Eq. (7)). The analysis of variance for the regression model (Eq. (7)) is given in Table 3. The determination of coefficient R2 = 0.835 indicates that 16.5% of total variation did not fit the model (Table 2). The statistical significance of the model equation was also confirmed by an F-test, where the value F = 3.109 exceeded the table value F (7, 6, 0.1) = 2.83. The Fisher F-test also indicates that the experimental results fitted the model well. Three-dimensional response plots are shown in Fig. 4. Increasing pretreatment time resulted in a significant decrease of methane yield; however, the methane yield increased when pretreatment time exceeded 21 days (Fig. 4a). A decrease of methane yield also occurred at a lower or higher pretreatment
Fig. 3. Example of mass balance and energy production for 12th pretreating experiment (0.12 g Ca(OH)2/g SC, 28 days at 45 °C).
342
Y.-g. Liang et al. / Chemical Engineering Journal 217 (2013) 337–344
Table 3 ANOVA for model.
Regression Residual Total
Sum of squares
df
Mean square
F
F(7,6,0.1)
p value
5744.076 1137.279 6881.354
7 6 13
820.582 189.546
3.109
2.83
0.047
2 weeks of the non-oxidative lime pretreatment [11]. The lower methane yield for medium-term pretreatment may be ascribed to a higher level of cellulose and hemicellulose removal. A lower cellulose and hemicellulose recovery was found at 2–6 weeks for non-oxidative and oxidative conditions at 55 °C [11]. Considering the three factors of pretreatment, lime loading had the largest influence on methane yield (Fig. 4 and Eq. (7)). 3.4. Lime pretreatment improving methane production rate constant
temperature. This is ascribed to less lignin removal at low temperatures and whereas a higher hemicellulose removal at high temperatures. Methane yield would increase with a rise in lime loading (Fig. 4b). This is ascribed to a higher level of lignin removal accompanied by sufficient deacetylation. The result was similar to the work by Chang et al. [13], who reported that the yield of reducing sugar increased when the lime loading was increased from 0.02 to 0.15 g Ca(OH)2/g dry biomass. From Fig. 4c, it was concluded that a short or long pretreatment time resulted in a higher methane yield, because most of the delignification happened in the first
The modeling results of methane production from raw and pretreated SC using a first-order kinetic model are shown in Table 4. All the plots showed a good linearity, with coefficients of determination greater than 0.84. The methane production rate constants (k) ranged from 0.064 to 0.128 d1. This was very close to the results of the BMP assays of switchgrass pretreated with microwaves (0.081–0.134 d1) [25] and Jatropha curcas L (0.069–0.138 d1) [24], and the k values in this study were higher than the results (0.01195–0.01571 d1) from digestion of landfilled solid wastes
Fig. 4. Response surface for the interactive effect on methane yield through BMP test (interactive effect of pretreatment time and temperature (a), lime loading and temperature (b), lime loading and pretreatment time (c)).
Y.-g. Liang et al. / Chemical Engineering Journal 217 (2013) 337–344 Table 4 Methane production rate constants and recovery rate of raw and pretreated SC. Run
K (d1)
R2
B0a (mL/g)
Methane recovery efficiency (%)
Raw P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13
0.041 0.070 0.064 0.066 0.064 0.074 0.070 0.076 0.092 0.112 0.128 0.119 0.108 0.066
0.844 0.974 0.936 0.968 0.944 0.965 0.920 0.938 0.937 0.987 0.949 0.978 0.990 0.966
281.4 291.4 297.0 282.8 293.0 286.3 305.9 251.7 251.6 247.9 256.0 223.5 266.5 323.3
45.1 64.1 56.5 56.5 57.3 64.7 59.0 72.7 76.9 77.2 83.4 86.0 81.4 62.6
R2, coefficients of determination; Methane recovery efficiency = the experimental methane yield 100%/the theoretical methane yield. a Theoretical methane yield.
[21], owing to a smaller particle size being used. Particle size reduction increases available specific surface and decreases the degree of polymerization [30]. These factors were all attributed to improve the hydrolysis rate of raw and pretreated SC. Compared with the k value of raw SC, the k value of pretreated SC increased between 56% and 212%. The highest k value happened with treatment P10, reaching 0.128 d1. The methane recovery efficiency was higher (56.5–86%) for pretreated SC when compared with raw SC (45.1%) (Table 4).
3.5. Economic and environmental comparison The method of pretreatment has been considered as a key step in the production of biofuels from lignocelluloses [1]. However, pretreatment is also one of the most energy intensive steps in the process [9]. Optimization of energy consumption was essential for the use of lime as a pretreatment to anaerobic digestion for the process to be economically feasible; therefore, a basic energy balance was preformed and the result is shown in Table S1. The samples from lime pretreatment produced a higher amount of methane compared to the untreated or raw sample. However, an additional energy with the value of 3.72–49.69 kJ/g TS for pretreatment biomass would be required, including the energy of heating pretreatment mixture and heat loss. Input energy was much higher than the increased energies provided by methane production enhancement from pretreated solid residues, yielding net energy balance of 0.42 to 46.81 kJ/g TS of biomass (Table S1). This implies that the increased energy production due to pretreatment was lower than the amount of input additional energy from pretreatment operation. So the reduction of heat loss energy and an additional energy production from anaerobic digestion of pretreatment liquid would be required. When the pretreatment liquid was subjected to anaerobic digestion, the energy of increased methane from pretreatment for P4, P7 and P10 could meet the energy requirement of pretreatment (Table S1). Considering the energy balance, the pretreatment condition of high lime loading, low temperature and short pretreatment time was a preferential selection, and in this study, the optimization of pretreatment experiment was P10 with pretreatment conditions of 0.10 g Ca(OH)2/g dry mass, at 25 °C for 14 days. Nevertheless, this was only a very rough analysis of energy. For a complete calculation, it should be necessary to take into account investment costs, maintenance costs and integration in the whole wastewater treatment process [31]. It should be noted that the aim of anaerobic digestion of SC not only was to obtain the energy, but also was to treat effectively the solid wastes of SC.
343
The major effect of lime pretreatment was the partly removal of lignin and hemicellulose from smooth cordgrass, thus improving the reactivity of the remaining polysaccharides [1]. However alkali extraction, including lime pretreatment could result in the solubilization, redistribution and condensation of lignin and the increscent in the crystalline state of the cellulose [32,33]. These effects would partly counteract the positive effects of lignin removal and cellulose swelling from pretreatment [34]. And simultaneously the residue of Ca2+ from pretreatment and the solubilized lignin components often had an inhibitory effect for methanogens [35], so the fermentation or anaerobic digestion of pretreated liquid from lime pretreatment required a conditioning treatment. P13 reduced the chemical costs compared to P1–P12 lime pretreatment since lime was not added, and simultaneously the pretreatment liquid of P13 also reduced the need for neutralization and conditioning chemicals. 4. Conclusions Lime pretreatment resulted in a significant change in SC compositions. The extent of change was a function of lime loading, pretreatment time and temperature. Lime pretreatment resulted in 5.7–60.5% hemicellulose and 10.2–36.2% lignin reductions, and good cellulose preservation with a 91–98.7% recovery yield. Lime pretreatment increased methane yield by 122–180% and the methane production rate constant by 56–212% for pretreated samples compared with raw sample. The highest methane yield with 218.4 mL/g TS was obtained from pretreatment conditions of lime loading 0.12 g Ca(OH)2/g SC, for 28 days at 45 °C. Lime loading had the greatest influence on methane yield of all the three pretreatment factors. Considering the energy balance, the optimization of pretreatment conditions was 0.10 g Ca(OH)2/g dry mass, at 25 °C for 14 days. Based on this study, the pretreatment condition of high lime loading, low temperature and short pretreatment time was a preferential consideration for lime pretreatment. Acknowledgements The work was partially supported by the China Ministry of Environmental Protection (2012ZX07102-004), Shanghai Science Foundation of China (11075039) and the Introducing and Stabilizing talents of Anhui Agricultural University (WD2012-2). Dr. Liang thanks Ms. Zhao Hong-fen for constant support and understanding. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.cej.2012.11.135. References [1] B. Yang, C.E. Wyman, Pretreatment: the key to unlocking low-cost cellulosic ethanol, Biofuels Bioprod. Bioref. 2 (2008) 26–40. [2] S.G. Yang, J.H. Li, Z. Zheng, Z. Meng, Characterization of Spartina alterniflora as feedstock for anaerobic digestion, Biomass Bioenergy 33 (2009) 597–602. [3] S.Q. An, B.H. Gu, C.F. Zhou, Z.S. Wang, Z.F. Deng, Y.B. Zhi, H.L. Li, L. Chen, D.H. Yu, Y.H. Liu, Spartina invasion in China: implications for invasive species management and future research, Weed Res. 47 (2007) 183–191. [4] G.Y. Chen, Z. Zheng, S.G. Yang, C.X. Fang, X.X. Zou, J.B. Zhang, Improve conversion of Spartina alterniflora into biogas by co-digestion with cow feces, Fuel Process. Technol. 91 (2010) 1416–1421. [5] Y.G. Liang, Z. Zheng, R.M. Hua, X.Z. Luo, A preliminary study of simultaneous lime treatment and dry digestion of smooth cordgrass for biogas production, Chem. Eng. J. 174 (2011) 175–181. [6] Y. Zhao, W.J. Lu, H.T. Wang, Supercritical hydrolysis of cellulose for oligosaccharide production in combined technology, Chem. Eng. J. 150 (2009) 411–417. [7] H.T. Tan, K.T. Lee, Understanding the impact of ionic liquid pretreatment on biomass and enzymatic hydrolysis, Chem. Eng. J. 183 (2012) 448–458.
344
Y.-g. Liang et al. / Chemical Engineering Journal 217 (2013) 337–344
[8] R. Haruna, M.K. Danquah, Enzymatic hydrolysis of microalgal biomass for bioethanol production, Chem. Eng. J. 168 (2011) 1079–1084. [9] N. Mosier, C. Wyman, B. Dale, R. Elander, Y.Y. Lee, M. Holtzapple, M. Ladisch, Features of promising technologies for pretreatment of lignocellulosic biomass, Bioresour. Technol. 96 (2005) 673–686. [10] R. Sierra, M.T. Holtzapple, C.B. Granda, Long-term lime pretreatment of poplar wood, AIChE J. 57 (2011) 1320–1328. [11] S. Kim, M.T. Holtzapple, Lime pretreatment and enzymatic hydrolysis of corn stover, Bioresour. Technol. 96 (2005) 1994–2006. [12] R. Sierra, L.A. Garcia, M.T. Holtzapple, Selectivity and delignification kinetics for oxidative and nonoxidative lime pretreatment of poplar wood, part III: long-term, Biotechnol. Progr. 26 (2010) 1685–1694. [13] V.S. Chang, M. Nagwani, M.T. Holtzapple, Lime pretreatment of crop residues bagasse and wheat straw, Appl. Biochem. Biotechnol. 74 (1998) 135–159. [14] S.C. Rabelo, H. Carrere, R.M. Filho, A.C. Costa, Production of bioethanol, methane and heat from sugarcane bagasse in a biorefinery concept, Bioresour. Technol. 102 (2011) 7887–7895. [15] H. Rughoonundun, C. Granda, R. Mohee, M.T. Holtzapple, Effect of thermochemical pretreatment on sewage sludge and its impact on carboxylic acids production, Waste Manage. 30 (2010) 1614–1621. [16] A.A. Modenbach, S.E. Nokes, The use of high-solids loadings in biomass pretreatment—a review, Biotechnol. Bioeng. 109 (2012) 1430–1442. [17] K.T. Fang, D.K.J. Lin, P. Winker, Y. Zhang, Uniform design: theory and application, Technometrics 42 (2000) 237–248. [18] R. Sierra, C.B. Granda, M.T. Holtzapple, Lime pretreatment, in: J.R. Mielenz (Ed.), Biofuels Methods and Protocols, Humana Press–Springer Protocols, New York, 2009, pp. 115–124. [19] B. Hames, R. Ruiz, C. Scarlata, A. Sluiter, J. Sluiter, D. Templeton, Preparation of samples for compositional analysis, Laboratory Analytical Procedure, National Renewable Energy Laboratory, 2008. [20] J.M. Owens, D.P. Chynoweth, Biochemical methane potential of municipal solid waste (MSW) components, Water Sci. Technol. 27 (1993) 1–14. [21] M.S. Bilgili, A. Demir, G. Varank, Evaluation and modeling of biochemical methane potential (BMP) of landfilled solid waste: a pilot scale study, Bioresour. Technol. 100 (2009) 4976–4980. [22] APHA, Standard methods for the examination of water and wastewater, 19th ed., Washington, DC, 1995.
[23] P.J. Van Soest, Use of detergents in the analysis of fibrous feeds. II. A rapid method for the determination of fiber and lignin, J. Assoc. Off. Anal. Chem. 46 (1963) 829–835. [24] V.N. Gunaseelan, Biomass estimates, characteristics, biochemical methane potential, kinetics and energy flow from Jatropha curcus on dry lands, Biomass Bioenergy 33 (2009) 589–596. [25] D. Jackowiak, J.C. Frigon, T. Ribeiro, A. Pauss, S. Guiot, Enhancing solubilisation and methane production kinetic of switchgrass by microwave pretreatment, Bioresour. Technol. 102 (2011) 3535–3540. [26] H.B. Klinke, B.K. Ahring, A.S. Schmidt, A.B. Thomsen, Characterization of degradation products from alkaline wet oxidation of wheat straw, Bioresour. Technol. 82 (2002) 15–26. [27] V.S. Chang, M.T. Holtzapple, Fundamental factors affecting biomass enzymatic reactivity, Appl. Biochem. Biotechnol. 84–86 (2000) 5–37. [28] K. Grohmann, D.J. Mitchell, M.E. Himmel, B.E. Dale, H.A. Schroeder, The role of ester groups in resistance of plant cell wall polysaccharides to enzymatic hydrolysis, Appl. Biochem. Biotechnol. 20–21 (1989) 45–61. [29] P. Alvira, E. Tomás-Pejó, M. Ballesteros, M.J. Negro, Pretreatment technologies for an efficient bioethanol production process based on enzymatic hydrolysis: a review, Bioresour. Technol. 101 (2010) 4851–4861. [30] L.M. Palmowski, J.A. Müller, Influence of the size reduction of organic waste on their anaerobic digestion, Water Sci. Technol. 41 (2000) 155–162. [31] C. Bougrier, J.P. Delgenès, H. Carrère, Impacts of thermal pre-treatments on the semi-continuous anaerobic digestion of waste activated sludge, Biochem. Eng. J. 34 (2007) 20–27. [32] A.T.W.M. Hendriks, G. Zeeman, Pretreatments to enhance the digestibility of lignocellulosic biomass, Bioresour. Technol. 100 (2009) 10–18. [33] Y.G. Liang, Z. Zheng, X.Z. Luo, F.H. Guo, L.M. Wang, J.B. Zhang, Effect of mesophilic and thermophilic conditions on changes of physicochemical characteristics of smooth cordgrass via dry digestion process, Chem. Eng. J. 168 (2011) 544–552. [34] D. Gregg, J.N. Saddler, A techno-economic assessment of the pretreatment and fractionation steps of a biomass-to-ethanol process, Appl. Biochem. Biotechnol. 57–58 (1996) 711–727. [35] Y. Chen, J.J. Cheng, K.S. Creamer, Inhibition of anaerobic digestion process: a review, Bioresour. Technol. 99 (2008) 4044–4064.