- Email: [email protected]
Effect of ozonolysis pretreatment on enzymatic digestibility of wheat and rye straw
Bioresource Technology 100 (2009) 1608–1613
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Effect of ozonolysis pretreatment on enzymatic digestibility of wheat and rye straw Ma Teresa García-Cubero, Gerardo González-Benito, Irune Indacoechea, Mónica Coca, Silvia Bolado * Department of Chemical Engineering and Environmental Technology, University of Valladolid, Po Prado de la Magdalena s/n, 47011 Valladolid, Spain
a r t i c l e
i n f o
Article history: Received 29 April 2008 Received in revised form 2 September 2008 Accepted 3 September 2008 Available online 31 October 2008 Keywords: Lignocellulose Pretreatment Ozone Enzymatic hydrolysis Straw
a b s t r a c t Wheat and rye straws were pretreated with ozone to increase the enzymatic hydrolysis extent of potentially fermentable sugars. Through a 25–1 factorial design, this work studies the influence of five operating parameters (moisture content, particle size, ozone concentration, type of biomass and air/ozone flow rate) on ozonization pretreatment of straw in a fixed bed reactor under room conditions. The acid insoluble lignin content of the biomass was reduced in all experiments involving hemicellulose degradation. Near negligible losses of cellulose were observed. Enzymatic hydrolysis yields of up to 88.6% and 57% were obtained compared to 29% and 16% in non-ozonated wheat and rye straw respectively. Moisture content and type of biomass showed the most significant effects on ozonolysis. Additionally, ozonolysis experiments in basic medium with sodium hydroxide evidenced a reduction in solubilization and/or degradation of lignin and reliable cellulose and hemicellulose degradation. Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction Lignocellulosic materials, especially agricultural and forestry residues, offer potential as cheap and abundant feedstock for production of fuels and chemicals through biological means (Hendriks and Zeeman, 2008; Mussato et al., 2008). Amongst possible applications, this research focuses on enzymatic hydrolysis of the cellulose fraction to glucose to be fermented in order to obtain fuel-grade ethanol in future research. Replacing fossil fuels with other renewable and less polluting fuels is the target set out in the European White Paper ‘‘European transport policy for 2010: time to decide”, published in 2001 and subsequently endorsed by Directive 2003/30/CE that establishes a reference value of a 5.75% market share for biofuels in 2010. Likewise, other countries such as the USA, Canada, China, Argentina or Brazil have fixed compulsory blending values of bioethanol in gasoline. Nevertheless, despite efforts to cut the production cost of ethanol, it remains too high compared to petroleum derived fuels. To reduce this cost, one interesting possibility is the use of lignocellulosic materials as feedstock (Sánchez and Cardona, 2007). The main obstacle to producing ethanol economically from lignocellulosic materials is its low digestibility due to the close link between their components: cellulose, hemicellulose and lignin. It is well known that lignin content significantly impacts enzymatic hydrolysis of lignocellulosic biomass (Mosier et al., 2005). Different authors have shown that biomass digestibility is enhanced by decreasing its lignin and/or hemicellulose content (Mussato et al., * Corresponding author. Tel.: +34 983423958; fax: +34 983423013. E-mail address: [email protected] (S. Bolado). 0960-8524/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2008.09.012
2008; Zhu et al., 2008). Therefore some pretreatment is necessary to modify lignocellulosic structure, solubilize and/or degrade its different components, and improve biomass digestibility at the subsequent enzymatic hydrolysis stage (Hendriks and Zeeman, 2008; Mosier et al., 2005). The different alternatives tested for lignocellulose pretreatment involve the use of physical, chemical, physicochemical and/or biological methods, e.g. steam explosion, hot water extraction (Rosgaard et al., 2007), sulfuric acid, sodium hydroxide, hydrogen peroxide, peracetic acid, ozonolysis (Silverstein et al., 2007; Zhao et al., 2008) ammonia fiber explosion, AFEX (Sendich et al., 2008; Teymori et al., 2005) and wet oxidation (Hendriks and Zeeman, 2008; Sun and Cheng, 2002). This research studies ozonization as an adequate chemical pretreatment. Ozone is a powerful oxidant, is soluble in water and is readily available. Ozone applications have increased substantially both in number and diversity over the last two decades, and have been used for example in the treatment of ground and industrial wastewaters (Amat et al., 2005; Coca et al., 2005). In pulp bleaching in the paper industry, ozone has been widely used and has evidenced high delignification efficiency (Roncero et al., 2003; Shatalov and Pereira, 2008). In recent decades, ozonolysis pretreatment has shown its efficacy essentially degrading the lignin polymer but also slightly solubilizing hemicellulose content of lignocellulosic biomass (Quesada et al., 1999; Sun and Cheng, 2002). Ozone is highly reactive towards compounds incorporating conjugated double bonds and functional groups with high electron densities. Therefore, the moiety most likely to be oxidized in ozonization of lignocellulosic materials is lignin due to its high content of C@C bounds. Ozone
1609
M.T. García-Cubero et al. / Bioresource Technology 100 (2009) 1608–1613
attacks lignin releasing soluble compounds of less molecular weight, mainly organic acids such as formic and acetic acid which can result in a drop in pH from 6.5 to 2. The main advantages linked to this process are the lack of any degradation products which might interfere with subsequent hydrolysis or fermentation, and the reactions occurring at ambient temperature (Contreras, 2002). Silverstein et al. (2007) recently applied ozone pretreatment for conversion of cotton stalks to ethanol by continuously sparging ozone gas through a 10%(w/v) mixture of cotton stalks and water at 4 °C for 30, 60 and 90 min. Ozone did not cause the expected effect, possibly because of insufficient reaction time, low ozone concentration or contact method. Additional research needs to be conducted to optimize ozonolysis pretreatment operation conditions. The aim of this work was to determine the influence of process parameters on the ozonolysis pretreatment of grain straw to improve yields at the enzymatic hydrolysis stage. A pilot plant at lab scale with a fixed bed reactor was designed and built for the ozonization experiments. The raw materials considered for this study were rye and wheat straw, both by-products from grain crop. These materials were selected as they are widely produced in the region of Castilla y León (Spain) and present a considerably different percentage of lignin. During an initial trial, the effect of five variables likely to impact the grain straw ozonolysis process such as moisture, particle size, ozone concentration, type of straw and air/ozone flow was studied, applying a 25–1 design experiment. A second test was conducted to optimize moisture value, experimentally found to be the most sensitive variable for both essayed raw materials. Finally, the pH process variable was analyzed, again using a 25–1 design experiment. 2. Methods 2.1. Raw material Wheat and rye straws were kindly donated by the Castilla y León Institute of Technological Agriculture. They were ground in a blender, sieved to obtain two different sizes: <1 mm and 3– 5 mm and kept in an oven at 45 °C. The composition (%w/w) of both straws is shown in Table 1.
the iodometric method (Standard Methods for the Examination of Water and Wastewater, 1995). Reactor outlet gas flow was passed through a 2% KI solution to remove any unreacted ozone from the gas stream. The resulting ozone-treated substrate was dried in an oven at 45 °C, stored in a freezer and used for enzymatic hydrolysis and/or for composition analysis. To check the effect of alkali conditions, several samples were hydrated with 20% NaOH solution instead of water prior to the reactor charge. Operating conditions used in the different trials are shown in Table 2. 2.3. Enzymatic hydrolysis Hydrolysis tests for each sample were performed to determine the improvement in enzymatic saccharification under the different pretreatment conditions applied. Enzymatic hydrolysis was performed using a mixture of cellulase complex (NS50013) and b-glucosidase (NS50010), both enzymes kindly donated by Novozymes (Denmark). Enzyme reactions were performed with 0.3 g of freeze stored biomass suspended in 5 mL acetate buffer 0.1 M (pH 4.8), containing 10 FPU g1 and 10 CBU g1 of substrate (dry basis) at 50 °C for 48 h. Test flasks were shaken in a rotary incubator at 150 rpm. After hydrolysis, 600 lL samples were withdrawn, passed through a 0.22 lm filter and stored for analysis of sugars and other possible compounds (e.g. inhibitors). 2.4. Analytical methods Acid insoluble lignin, acid soluble lignin, cellulose and hemicellulose in the raw material were estimated following NREL laboratory analytical procedures Lap 003, 004 and 002 respectively (NREL, 1995), with the exception of a Bio-Rad HPX-87 H ionexclusion column used to measure sugar concentration. The mobile phase was 5 mM H2SO4 at a flow rate of 0.6 mL min1 at 60 °C. The detector was based on the refraction index measurement. Sugars and other possible inhibitors (HMF, furfural, etc.) from enzymatic hydrolysis were also analyzed by HPLC using the Aminex HPX-87 H column (Bio-Rad, Hercules, CA) under the same operating conditions as indicated previously. Neutral detergent fiber (NDF) and acid detergent fiber (ADF) were prepared following Goering and Van Soest (1970) to determine cellulose and hemicellulose fraction.
2.2. Ozonolysis pretreatment 2.5. Statistical analysis The ozonolysis pretreatment of wheat and rye straw was performed in a fixed bed reactor (glass column 50 cm in height and 2.7 cm in diameter) under room conditions. At the beginning of each test, ground material was hydrated to the required value, fed into the reactor until total reaction volume was attained and then exposed to an air/ozone gas stream in the fixed bed reactor. Kinetic experiments were carried out to determine a reaction time long enough to allow total oxidation of the straw. Ozone production (ozonizator Sander 301) was controlled by varying either the air flow rate or the electrical power supply. Ozone concentration in the gas phase was measured following
Table 1 Composition of wheat and rye straw Composition
Wheat, % (w/w)
Rye, % (w/w)
Moisture Cellulose (as glucose) Hemicellulose (as xylose) Acid insoluble lignin Acid soluble lignin
9.0 34.2 20.1 17.1 5.2
5.9 30.9 21.5 22.1 3.2
The experimental work was conducted in three consecutive trials. The first trial was designed to examine the influence of the ozone pretreatment operation factors [Moisture (M), particle size (S), ozone concentration (O), type of biomass (L) and air/O3 flow rate (C)] on straw enzymatic digestibility. Experiments were
Table 2 Experimental levels of the parameters studied in the experimental design for trial s 1 and 3 Factor
Moisture (M) (%w/w) Particle size (S) (mm) Ozone concentration (O) (%w/w) Type of biomass (L) Air/ozone flow rate (C) (L/ha) Basic conditions (B) (% NaOHb) a
Level value Lower (1)
Upper (1)
20 <1 2.7 Wheat 60 0
40 3–5 3 Rye 90 20
Operation variable in trial 1. Operation variable in trial 3. When parameter B is in the upper level, samples were moistened with 20% NaOH solution. The air/ozone flow in this trial experiments was 60 L/h. b
1610
M.T. García-Cubero et al. / Bioresource Technology 100 (2009) 1608–1613
analyzed through a 25–1 factorial design in which the variables were evaluated at two levels (Table 2). Assessing the impact of the main operating factors and their interactions was calculated and laid out in a normal probability plot to draw the most significant factors of the process. Estimating the effect is defined as the change in the response caused by a change in the level of the parameter. The second trial was performed to examine the effect of moisture content during ozone treatment. Water was added at several water-to-straw ratios to obtain materials with different moisture content. Finally, a third trial was conducted to determine the effect of basic conditions on ozonolysis pretreatment by moistening the samples with 20% NaOH. Basic impregnated wheat and rye straw chips were pretreated under eight distinct experimental conditions. These conditions were derived from the same factorial design as in the first trial with a change in parameter air/O3 flow rate (C) for which a value of 60 L/h for pH condition (B) (Table 2) was maintained in this trial.
3. Results and discussion 3.1. Determination of ozonolysis time To set the time required for the reaction, several ozonization tests were run modifying reaction time for rye straw. All these experiments were carried out with an air/ozone flow of 60 L/h, 2.7% w/w ozone, 40% w/w moisture and 3–5 mm particle size. The most striking ozone pretreatment effects observed were acid insoluble and soluble lignin variations. Evolution of these variables was followed to establish reaction time in further experiments. Results obtained showed that ozonization effects remained essentially constant for reaction times above 2.5 h. Acid insoluble lignin (AIL) content decreased from 22.1% (w/w) for non-ozonated material to 12.3% at ozonation times above 2.5 h, whereas acid soluble lignin (ASL) content increased from 3.2% for raw material to 8.1% after 2.5 h. Ozonization pretreatment seems to solubilize more than to degrade the initial lignin present in raw material. From these results a reaction time of 2.5 h was selected for subsequent trials. 3.2. Effect of ozone pretreatment operation variables on straw delignification As explained above, the influence of five parameters on a 25–1 (16 run) factorial design was assessed. This method was used to provide a good insight into pretreatment optimization, establishing which parameters have a significant effect on the ozonization of wheat and rye straw, either directly or by their interactions. Possible inhibitors such as furfural or HMF were not detected in ozonated samples. The impact of ozone on lignin content under the different operating conditions is illustrated in Table 3. All experiments led to a reduction in acid insoluble lignin compared to untreated biomass. ASL content increases with ozone pretreatment compared to both untreated straws, although no clear relationship between ozonization conditions and lignin solubilization was found. Possible structural changes arising from ozone pretreatment were studied analysing the contents of Acid Detergent Fiber (ADF) and Neutral Detergent Fiber (NDF). Results for wheat and rye straw before and after ozonolysis pretreatment of samples under the different operating conditions from the first trial are presented in Fig. 1 as a function of AIL variation in the pretreated samples. The continuous lines in each figure represent the theoretical value that ADF (comprising cellulose and lignin fraction) and NDF (comprising cellulose, hemicellulose and lignin fractions)
Table 3 Results obtained in the 25-1 experimental design for trial 1
msolc MsoLc mSoLc MSolc msoLC MsolC mSolC MSoLC msOLc MsOlc mSOlc MSOLc msOlC MsOLC mSOLC MSO1C
%AIL (%w/w)
%ASL (%w/w)
AIL variation (%w/w)
Effect estimation Parameter
AIL variation
11,8 13,2 16,2 13,5 18,3 12,2 13,5 12,1 17,7 11,8 14,4 11,7 13,9 12,4 17,1 11,2
9,3 9,3 7,2 7,3 6,3 10,4 8,6 7,4 7,7 8,6 8,5 8,5 9,7 8,9 6,6 8,4
32,4 40,3 26,5 22,7 17,1 30,1 22,7 45,4 19,8 32,4 17,7 47,0 20,6 43,8 22,6 35,9
M S c o L CO SO MO OL SC MC CL MS SL ML MSCOL
59,2 2,0 0,6 1,3 24,0 6,2 4,8 19,3 2,5 12,9 13,1 4,4 2,4 18,5 31,3 238,5
M, S, O, L and C are the five considered parameters, e.g. moisture, particle size, ozone concentration, type of biomass and air/ozone flow rate. Upper case letters and lower case letters were respectively used to represent upper and lower level of every considered parameter. %AIL is the acid insoluble lignin obtained for every pretreated samples, %ASL is the acid soluble lignin, AIL variation is the amount of acid insoluble lignin regarding the AIL presented in raw materials and Effect estimation is the change in the response that is produced by a change in the level of the parameter, e.g. acid insoluble lignin variation.
a
wheat straw
100 80
60 % NDF % ADF 40 20 0
b
0
10 20 30 % Acid Insoluble Lignin
40
rye straw
100 80
% NDF 60 % ADF 40 20 0
0
10
20 30 40 %Acid Insoluble Lignin
50
Fig. 1. NDF and ADF as a function of the percentage of Acid Insoluble Lignin removed by the ozonolysis pretreatment of a) wheat straw and b) rye straw. (h)% NDF for wheat straw. (s)% ADF for wheat straw. (e)% NDF for rye straw. (4)% ADF for rye straw. The continuous line indicates the theoretical values if only lignin was degraded. The dotted line shows the linear trend for the experimental points.
should have when eliminating only the lignin fraction calculated from results in Table 3. The dotted row is the line followed by the real points of ADF and NDF obtained in sample analysis. In Fig. 1a), the continuous and dotted lines for ADF wheat straw
1611
M.T. García-Cubero et al. / Bioresource Technology 100 (2009) 1608–1613
overlap, meaning that the theoretical and experimental ADF values are similar. It may thus be inferred from Fig. 1a) that there was no degradation of cellulose due to ozone pretreatment for wheat straw. However, in Fig. 1b) real ADF values for rye straw deviate slightly from the theoretical line, although only in samples with high acid insoluble lignin removal yield. Results thus evidence a small decrease in cellulose (up to 5%) for lignin removals up to 35%, indicating that ozone is not totally selective with lignin. Fig. 1 shows how experimental NDF values deviate from theoretical points for both raw materials. Taking into account that there are no cellulose losses in wheat straw and since the loss of this fraction in rye straw is below 5%, differences between real and theoretical values obtained should be due to hemicellulose solubilization. The higher difference between the dotted and continuous lines corresponds to the samples with a high degree of delignification. Therefore, solubilization of hemicellulose and lignin is likely to occur at the same time. These experimental results were confirmed by chromatographic analysis in some of the pretreated samples (data not shown). This behaviour has also been evidenced by other authors. Binder et al. (1980) and Ghedalia and Miron (1981) observed that losses in cellulose fractions due to pretreatment were below 5% when working at low pH values. Ghedalia and Miron (1981) pointed out that ozone not only led to a decrease in lignin content but also to the solubilization of the hemicellulose fraction converting it into cell solubles. Sun and Cheng (2002) indicated that as lignin content fell, solubilization of hemicellulose increased, likely because both compounds solubilized together in a lignin-hemicellulose complex. From the previous results, acid insoluble lignin concentration was therefore selected as the variable to study the effect of ozone pretreatment on straw digestibility. For each of the 16 runs, the percentage of AIL variation and its main effects and interactions are given. The percentage of AIL variation was calculated taking into account that the percentage of acid insoluble lignin was 17.1 and 22.1 for untreated wheat and rye straw respectively. Statistical treatment of the results in Table 3 revealed that moisture and biomass type (L) were the most significant variables from the studied parameters, always working with an identical operation time (2.5 h). Fig. 2 represents the normal probability plot, in which these effects deviate from the line that goes through the remaining points. The value of the ‘‘biomass type” (L) effect was positive in the normal probability plot, ozonization pretreatment thus removing a higher percentage of acid insoluble lignin in rye than in wheat
straw, although the lowest remaining lignin obtained in the pretreated samples was analogous for both materials (around 11– 12%). A fraction of acid insoluble lignin exposed to ozone can probably be solubilized or degraded, leaving cellulose more accessible to enzymes as suggested by Neely (1984), although there must be another fraction of lignin refractory to ozonization pretreatment. As before, following the factorial design, as the percentage of water content increased, an increase in the removal of acid insoluble lignin for both type of biomass was observed. The remaining parameters studied evidenced no major significance from the point of view of delignification yield and digestibility, although they do need to be taken into account to make this pretreatment feasible from an economical standpoint. Results from the first trial allow us to select standard conditions for the parameters considered to be used in all subsequent experiments, e.g. air/ozone flow rate of 60 L/h, 2.7% w/w ozone concentration and a 3–5 mm particle size. Experiments in the first trial were carried out with two different moisture contents. In order to evaluate the possibility of a non-lineal relationship between delignification and moisture, seven runs with different water content were performed for both straws. Experimental results are shown in Fig. 3. AIL solubilization behaviour during ozonolysis pretreatment as moisture content increases is very similar for both straws. Fig. 3 shows how AIL content diminished as moisture increased up to 30%, likely due to an improvement in mass transfer from air/ozone flow rate to solid surface. An increase in water content above 30% evidenced no significant effects on ozonolysis pretreatment. Working with oak sawdust, Neely (1984) states the optimum range for water content should be 25–35% by weight, and Vidal and Molinier (1988) working with poplar sawdust obtained an optimum water content of 75%, probably due to the higher ozone flow rates employed in the tests. Working with cotton stalks with much higher water content at a solid loading of 10% w/v, Silverstein et al. (2007) obtained no significant changes in lignin, xylan or glucan contents. To study the influence of basic conditions on ozonolysis pretreatment, pH has been included in the factorial design instead of air/O3 flow rate (see Table 2). For statistical interpretation of this influence, the percentage of AIL variation in each sample and the estimation of its principal effects and their interactions are presented in Table 4. Fig. 4 shows the most important factors in ozonolysis pretreatment in basic conditions to be pH and, to a lesser extent, moisture and biomass type. The negative value of parameter B indicates that when basic conditions were used the percentage of acid insoluble lignin removed decreased. Lignin removal ranged between 1.6 and 19.6%
4 3
20
L 1 0 -40
-20
-1
0
20
40
60
80
-2
% Acid Lignin
M
2
15 10 5 0 0
-3 Fig. 2. Normal probability plot of the effects of parameters in the experimental design (trial 1) for AIL variation. X-axis represents the effect estimation and Y-axis the normal probability (depends on the number of experimental runs).
20
40 60 Water Content (%)
80
Fig. 3. Effect of moisture during ozonolysis pretreatment for both straws. Study of the amount of AIL and ASL for each moisture content. (h)% AIL for wheat straw. (s)% ASL for wheat straw. (e)% AIL for rye straw. (4)% ASL for rye straw.
M.T. García-Cubero et al. / Bioresource Technology 100 (2009) 1608–1613
Table 4 Results obtained in the 25–1 experimental design (trial 3)
Msolb MsoLb mSoLb Msolb msoLB MsolB mSolB MSoLB msOLb MsOlb mSOlb MSOLb msOlB MsOLB mSOLB MSO1B
100
%AIL
%ASL
AIL variation
Effect estimation
(%w/w)
(%w/w)
(%w/w)
Parameter
AIL variation
11,8 13,2 16,3 13,5 18,7 17,0 14,8 17,1 17,7 11,8 14,4 11,7 15,9 16,5 19,1 14,1
9,3 9,3 7,2 7,3 5,9 7,1 7,0 6,6 7,7 8,6 8,5 8,5 7,8 8,1 6,6 6,5
32,4 40,3 26,5 22,7 15,6 1,6 15,6 22,6 19,8 32,4 17,7 47,0 9,3 9,7 13,5 19,6
M S B 0 L OB SO MO OL SB MB LB MS SL ML MSOLB
22,8 12,1 65,6 4,1 21,9 0,8 14,6 25,6 10,9 23,1 23,3 6,6 15,7 12,2 21,5 173,2
a B correspond with the parameter ‘‘Basic conditions”. Those experiences which were moistened with 20% NaOH solution.
2
MB
1 0 -90
-40
10
60
-1
B
-2
Fig. 4. Normal probability plot of the effects of the parameters in the experimental design (trial 3) for AIL variation. X-axis represents the effect estimation and Y-axis the normal probability (depends on the number of experimental runs).
for wheat straw and 9.7–22.6% for rye straw, lower than results derived from the first trial. Statistical analysis showed a lower influence of moisture and biomass type, probably due to the significant impact of the pH parameter. Alkaline conditions for ozonolysis pretreatment were used by Binder et al. (1980) in a stirred semibatch reactor. However, the delignification extent obtained diminished noticeably with acid insoluble lignin compositions near to 16– 17% being obtained. Additionally, results from NDF and ADF analysis of ozone pretreated straw in the third trial, when sodium hydroxide was added, showed an increase in cellulose degradation of up to 40% for rye straw and 50% for wheat straw, whereas 50% of the hemicellulose present in both straws was degraded. These results agree with those obtained by Binder et al. (1980), Roncero et al. (2001) and Coca et al. (2005). These authors conclude that under basic conditions hydroxide ions catalyze the decomposition of ozone to yield highly reactive and non-selective hydroxyl radicals, delignification decreases, and degradation of cellulose and hemicellulose increases. 3.3. Enzymatic hydrolysis The yield obtained after enzymatic hydrolysis of the pretreated samples from the first and second trials, defined as the percentage of glucose hydrolyzed, was correlated with the acid insoluble lig-
% Glucose hydrolyzed
1612
80 60 40 20 0
10
15 20 % Acid Insoluble Lignin
25
Fig. 5. Enzymatic hydrolysis yield as a function of the percentage of acid insoluble lignin present in the pretreated samples for wheat straw (h) and rye straw (e).
nin concentration as shown in Fig. 5, for wheat and rye straw respectively, in agreement with results obtained by Zhu et al. (2008) when poplar wood was pretreated and hydrolyzed. The ozonated material from the third trial was not enzymatically hydrolyzed due to the high losses observed in cellulose fraction after ozone pretreatment at high pH. In all cases, when the ozonated materials were hydrolyzed, the yield obtained increased compared to untreated straw. From the two types of straws compared, wheat proved easier to hydrolyze than rye, even when not being treated with ozone. Enzymatic hydrolysis yield improved from 29% corresponding to untreated raw material to 53–88.6% for pretreated wheat straw, corresponding to extreme values of acid insoluble lignin content obtained in ozonated samples, which decreased from 17.1% to 13.5% and 11.2%, respectively. The same behaviour was observed for rye straw. The enzymatic hydrolysis yield improved from 16% corresponding to untreated material to 36–57% for pretreated samples when acid insoluble lignin content was reduced from 22.1% to 18.3% and 12.1%, respectively. Therefore, ozonolysis pretreatment partially removes the acid insoluble lignin content of the biomass, increasing the yield in subsequent enzymatic hydrolysis. The enzymatic hydrolysis yields obtained with ozone pretreated straw in this work are in the same range as the best results found for these raw materials. Sun and Chen (2008) found 90% enzymatic hydrolysis yield when wheat straw was pretreated by aqueous glycerol, with a 70% removal of hemicellulose and 65% of lignin. García-Aparicio et al. (2007), working with steam exploded barley straw (210 °C, 10 min), reported values between 58.2% and 85.7% after hydrolysis times of 24 and 120 h respectively. Rosgaard et al. (2007) obtained lower glucose yields (39% w/w) with pretreated wheat straw using hot water extraction (16% dry matter, 60 °C, 15 min; liquids removed; 180 °C, 10 min; 195 °C, 3 min). In the same work, water impregnated and steam exploded wheat straw (70% dry matter, 1 h, 220 °C and 2.5 min) showed a glucose yield of 30% after 72 h of enzymatic hydrolysis. 4. Conclusions Ozonolysis is an efficient pretreatment for cereal straw. Ozone degrades and/or solubilizes lignin and slightly solubilizes the hemicellulose fraction, improving subsequent enzymatic hydrolysis. Negligible losses of cellulose were detected. Operating in a fixed bed reactor, moisture and biomass type proved the most relevant parameters. Moisture is a reaction controlling parameter for values below 30%. Wheat straw proved easier to hydrolyze than rye, although a similar content of residual lignin after ozone pretreatment was obtained for both essayed materials.
M.T. García-Cubero et al. / Bioresource Technology 100 (2009) 1608–1613
Ozone pretreatment in a basic medium did not prove adequate. Addition of sodium hydroxide reduces the delignification effect and increases degradation of both cellulose and hemicellulose fractions. Acknowledgements The authors wish to thank the Castilla y León Institute of Technological Agriculture (Ref. VA-01/2005), the Regional Ministry of Education at the Castilla y León Regional Government (VA060A06) and the Spanish Ministry of Education and Science (CTQ2006-15217/ PPQ) for financially supporting this research. Irune Indacoechea is also grateful to the Castilla y León Regional Government for providing her Doctorate Scholarship. References Amat, A.M., Arques, A., Miranda, M.A., López, F., 2005. Use of ozone and/or UV in the treatment of effluents from board paper industry. Chemosphere 60, 1111–1117. Binder, A., Pelloni, L., Fiechter, A., 1980. Delignification of straw with ozone to enhance biodegradability. Eur. J. Appl. Microbiol. Biotechnol. 11, 1–5. Coca, M., Peña, M., González, G., 2005. Variables affecting efficiency of molasses fermentation wastewater ozonization. Chemosphere 60, 1408–1415. Contreras, S., 2002. Degradation and biodegradability enhancement of nitrobenzene and 2,4-dichlorophenol by means of advanced oxidation processes based on ozone. PhD Thesis. University of Barcelona. 84-688–2054-7. García-Aparicio, M.P., Ballesteros, M., Manzanares, P., Ballesteros, I., González, A., Negro, M.J., 2007. Xylanase contribution to the efficiency of cellulose enzymatic hydrolysis of barley straw. Appl. Biochem. Biotechnol. 136–140, 353–365. Ghedalia, D.B., Miron, J., 1981. The effect of combined chemical and enzyme treatments on the saccharification and in vitro digestion rate of wheat straw. Biotechnol. Bioeng. 23, 823–831. Goering, H.D., Van Soest, P.J., 1970. Forage fiber analyses. Agric. Handbook., N 379. USDA, Washington DC. Hendriks, A.T.W.M., Zeeman, G., 2008. Pretreatments to enhance the digestibility of lignocellulosic biomass. Bioresour. Technol. doi:10.1016/j.biortech.2008.05.027. Mosier, N., Wyman, C., Dale, B., Elander, R., Lee, Y.Y., Holtzapple, M., Ladish, M., 2005. Features of promising technologies for pretreatment of lignocellulosic biomass. Bioresour. Technol. 96, 673–686. Mussato, S.I., Fernandes, M., Milagres, A.M.F., Roberto, I.C., 2008. Effect of hemicellulose and lignin on enzymatic hydrolysis of cellulose from brewer‘s spent grain. Enzyme Microb. Technol. 34, 124–129.
1613
National Renewable Energy Laboratory (NREL), 1995. Chemical Analysis and Testing Laboratory Analytical Procedures: LAP-002, LAP-003, LAP-004. NREL, Golden, CO, USA.. Neely, W.C., 1984. Factors affecting the pretreatment of biomass with gaseous ozone. Biotechnol. Bioenergy 26, 59–65. Quesada, J., Rubio, M., Gómez, D., 1999. Ozonation of lignin rich solid fractions from corn stalks. J. Wood Chem. Tech. 19, 115–137. Roncero, M.B., Colom, J.F., Vidal, T., 2001. Incremento de la selectividad en el blanqueo con ozono de una Secuencia TCF del tipo XOZP. Parte II. Aplicación de aditivos. Afinidad 58 (492), 121–128. Roncero, M.B., Torres, A.L., Colom, J.F., Vidal, T., 2003. TCF bleaching of wheat straw pulp using ozone and xylanase. Part A: paper quality assessment. Bioresour. Technol. 87, 305–314. Rosgaard, L., Pedersen, S., Meyer, A.S., 2007. Comparison of different pretreatment strategies for enzymatic hydrolysis of wheat and barley straw. Appl. Biochem. Biotechnol. 143, 284–296. Sánchez, O.J., Cardona, C.A., 2007. Trends in biotechnical production of fuel ethanol from different feedstocks. Bioresour. Technol. 99, 5270–5295. Sendich, E., Laser, M., Kim, S., Alizadeh, H., Laureano-Perez, L., Dale, B., Lynd, L., 2008. Recent progress improvements for the ammonia fiber explosion (AFEX) process and resulting reductions in minimum ethanol selling price. Bioresour. Technol. doi:10.1016/j.biortech.2008.02.059. Shatalov, A.A., Pereira, H., 2008. Arundo donax L. reed: new perspectives for pulping and bleaching. 5. Ozone-based TCF bleaching and organosolv pulps. Bioresour. Technol. 99, 472–478. Silverstein, R.A., Chen, Y., Sharma-Shivappa, R.R., Boyette, M.D., Osborne, J., 2007. A comparison of chemical pretreatment methods for improving saccharification of cotton stalks. Bioresour. Technol. 98, 3000–3011. Standard Methods for the Examination of Water and Wastewater, 1995, 19th ed. APHA-AWWA-WEF, Washington DC., USA. Sun, F., Chen, H., 2008. Enhanced hydrolysis of wheat straw by aqueous glycerol pretreatment. Bioresour. Technol. doi:10.1016/j.biortech.2007.12.027. Sun, Y., Cheng, J., 2002. Hydrolysis of lignocellulosic materials for ethanol production: a review. Bioresour. Technol. 83, 1–11. Teymori, F., Laureano-Perez, L., Alizadeh, H., Dale, B.E., 2005. Optimization of the ammonia fiber explosion (AFEX) treatment parameters for enzymatic hydrolysis of corn stover. Bioresour. Technol. 96, 2014–2018. Vidal, P.F., Molinier, J., 1988. Ozonolysis of lignin-improvement of in vitro digestibility of poplar sawdust. Biomass 16, 1–17. Zhao, X., Zhang, L., Liu, D., 2008. Comparative study on chemical pretreatment methods for improving enzymatic digestibility of croton weed stem. Bioresour. Technol. 99, 3729–3736. Zhu, L., O’Dwyer, J.P., Chang, V.S., Granda, C.B., Holtzapple, M.T., 2008. Structural features affecting biomass enzymatic digestibility. Bioresour. Technol. 99, 3817–3828.
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Effect of ozonolysis pretreatment on enzymatic digestibility of wheat and rye straw Ma Teresa García-Cubero, Gerardo González-Benito, Irune Indacoechea, Mónica Coca, Silvia Bolado * Department of Chemical Engineering and Environmental Technology, University of Valladolid, Po Prado de la Magdalena s/n, 47011 Valladolid, Spain
a r t i c l e
i n f o
Article history: Received 29 April 2008 Received in revised form 2 September 2008 Accepted 3 September 2008 Available online 31 October 2008 Keywords: Lignocellulose Pretreatment Ozone Enzymatic hydrolysis Straw
a b s t r a c t Wheat and rye straws were pretreated with ozone to increase the enzymatic hydrolysis extent of potentially fermentable sugars. Through a 25–1 factorial design, this work studies the influence of five operating parameters (moisture content, particle size, ozone concentration, type of biomass and air/ozone flow rate) on ozonization pretreatment of straw in a fixed bed reactor under room conditions. The acid insoluble lignin content of the biomass was reduced in all experiments involving hemicellulose degradation. Near negligible losses of cellulose were observed. Enzymatic hydrolysis yields of up to 88.6% and 57% were obtained compared to 29% and 16% in non-ozonated wheat and rye straw respectively. Moisture content and type of biomass showed the most significant effects on ozonolysis. Additionally, ozonolysis experiments in basic medium with sodium hydroxide evidenced a reduction in solubilization and/or degradation of lignin and reliable cellulose and hemicellulose degradation. Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction Lignocellulosic materials, especially agricultural and forestry residues, offer potential as cheap and abundant feedstock for production of fuels and chemicals through biological means (Hendriks and Zeeman, 2008; Mussato et al., 2008). Amongst possible applications, this research focuses on enzymatic hydrolysis of the cellulose fraction to glucose to be fermented in order to obtain fuel-grade ethanol in future research. Replacing fossil fuels with other renewable and less polluting fuels is the target set out in the European White Paper ‘‘European transport policy for 2010: time to decide”, published in 2001 and subsequently endorsed by Directive 2003/30/CE that establishes a reference value of a 5.75% market share for biofuels in 2010. Likewise, other countries such as the USA, Canada, China, Argentina or Brazil have fixed compulsory blending values of bioethanol in gasoline. Nevertheless, despite efforts to cut the production cost of ethanol, it remains too high compared to petroleum derived fuels. To reduce this cost, one interesting possibility is the use of lignocellulosic materials as feedstock (Sánchez and Cardona, 2007). The main obstacle to producing ethanol economically from lignocellulosic materials is its low digestibility due to the close link between their components: cellulose, hemicellulose and lignin. It is well known that lignin content significantly impacts enzymatic hydrolysis of lignocellulosic biomass (Mosier et al., 2005). Different authors have shown that biomass digestibility is enhanced by decreasing its lignin and/or hemicellulose content (Mussato et al., * Corresponding author. Tel.: +34 983423958; fax: +34 983423013. E-mail address: [email protected] (S. Bolado). 0960-8524/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2008.09.012
2008; Zhu et al., 2008). Therefore some pretreatment is necessary to modify lignocellulosic structure, solubilize and/or degrade its different components, and improve biomass digestibility at the subsequent enzymatic hydrolysis stage (Hendriks and Zeeman, 2008; Mosier et al., 2005). The different alternatives tested for lignocellulose pretreatment involve the use of physical, chemical, physicochemical and/or biological methods, e.g. steam explosion, hot water extraction (Rosgaard et al., 2007), sulfuric acid, sodium hydroxide, hydrogen peroxide, peracetic acid, ozonolysis (Silverstein et al., 2007; Zhao et al., 2008) ammonia fiber explosion, AFEX (Sendich et al., 2008; Teymori et al., 2005) and wet oxidation (Hendriks and Zeeman, 2008; Sun and Cheng, 2002). This research studies ozonization as an adequate chemical pretreatment. Ozone is a powerful oxidant, is soluble in water and is readily available. Ozone applications have increased substantially both in number and diversity over the last two decades, and have been used for example in the treatment of ground and industrial wastewaters (Amat et al., 2005; Coca et al., 2005). In pulp bleaching in the paper industry, ozone has been widely used and has evidenced high delignification efficiency (Roncero et al., 2003; Shatalov and Pereira, 2008). In recent decades, ozonolysis pretreatment has shown its efficacy essentially degrading the lignin polymer but also slightly solubilizing hemicellulose content of lignocellulosic biomass (Quesada et al., 1999; Sun and Cheng, 2002). Ozone is highly reactive towards compounds incorporating conjugated double bonds and functional groups with high electron densities. Therefore, the moiety most likely to be oxidized in ozonization of lignocellulosic materials is lignin due to its high content of C@C bounds. Ozone
1609
M.T. García-Cubero et al. / Bioresource Technology 100 (2009) 1608–1613
attacks lignin releasing soluble compounds of less molecular weight, mainly organic acids such as formic and acetic acid which can result in a drop in pH from 6.5 to 2. The main advantages linked to this process are the lack of any degradation products which might interfere with subsequent hydrolysis or fermentation, and the reactions occurring at ambient temperature (Contreras, 2002). Silverstein et al. (2007) recently applied ozone pretreatment for conversion of cotton stalks to ethanol by continuously sparging ozone gas through a 10%(w/v) mixture of cotton stalks and water at 4 °C for 30, 60 and 90 min. Ozone did not cause the expected effect, possibly because of insufficient reaction time, low ozone concentration or contact method. Additional research needs to be conducted to optimize ozonolysis pretreatment operation conditions. The aim of this work was to determine the influence of process parameters on the ozonolysis pretreatment of grain straw to improve yields at the enzymatic hydrolysis stage. A pilot plant at lab scale with a fixed bed reactor was designed and built for the ozonization experiments. The raw materials considered for this study were rye and wheat straw, both by-products from grain crop. These materials were selected as they are widely produced in the region of Castilla y León (Spain) and present a considerably different percentage of lignin. During an initial trial, the effect of five variables likely to impact the grain straw ozonolysis process such as moisture, particle size, ozone concentration, type of straw and air/ozone flow was studied, applying a 25–1 design experiment. A second test was conducted to optimize moisture value, experimentally found to be the most sensitive variable for both essayed raw materials. Finally, the pH process variable was analyzed, again using a 25–1 design experiment. 2. Methods 2.1. Raw material Wheat and rye straws were kindly donated by the Castilla y León Institute of Technological Agriculture. They were ground in a blender, sieved to obtain two different sizes: <1 mm and 3– 5 mm and kept in an oven at 45 °C. The composition (%w/w) of both straws is shown in Table 1.
the iodometric method (Standard Methods for the Examination of Water and Wastewater, 1995). Reactor outlet gas flow was passed through a 2% KI solution to remove any unreacted ozone from the gas stream. The resulting ozone-treated substrate was dried in an oven at 45 °C, stored in a freezer and used for enzymatic hydrolysis and/or for composition analysis. To check the effect of alkali conditions, several samples were hydrated with 20% NaOH solution instead of water prior to the reactor charge. Operating conditions used in the different trials are shown in Table 2. 2.3. Enzymatic hydrolysis Hydrolysis tests for each sample were performed to determine the improvement in enzymatic saccharification under the different pretreatment conditions applied. Enzymatic hydrolysis was performed using a mixture of cellulase complex (NS50013) and b-glucosidase (NS50010), both enzymes kindly donated by Novozymes (Denmark). Enzyme reactions were performed with 0.3 g of freeze stored biomass suspended in 5 mL acetate buffer 0.1 M (pH 4.8), containing 10 FPU g1 and 10 CBU g1 of substrate (dry basis) at 50 °C for 48 h. Test flasks were shaken in a rotary incubator at 150 rpm. After hydrolysis, 600 lL samples were withdrawn, passed through a 0.22 lm filter and stored for analysis of sugars and other possible compounds (e.g. inhibitors). 2.4. Analytical methods Acid insoluble lignin, acid soluble lignin, cellulose and hemicellulose in the raw material were estimated following NREL laboratory analytical procedures Lap 003, 004 and 002 respectively (NREL, 1995), with the exception of a Bio-Rad HPX-87 H ionexclusion column used to measure sugar concentration. The mobile phase was 5 mM H2SO4 at a flow rate of 0.6 mL min1 at 60 °C. The detector was based on the refraction index measurement. Sugars and other possible inhibitors (HMF, furfural, etc.) from enzymatic hydrolysis were also analyzed by HPLC using the Aminex HPX-87 H column (Bio-Rad, Hercules, CA) under the same operating conditions as indicated previously. Neutral detergent fiber (NDF) and acid detergent fiber (ADF) were prepared following Goering and Van Soest (1970) to determine cellulose and hemicellulose fraction.
2.2. Ozonolysis pretreatment 2.5. Statistical analysis The ozonolysis pretreatment of wheat and rye straw was performed in a fixed bed reactor (glass column 50 cm in height and 2.7 cm in diameter) under room conditions. At the beginning of each test, ground material was hydrated to the required value, fed into the reactor until total reaction volume was attained and then exposed to an air/ozone gas stream in the fixed bed reactor. Kinetic experiments were carried out to determine a reaction time long enough to allow total oxidation of the straw. Ozone production (ozonizator Sander 301) was controlled by varying either the air flow rate or the electrical power supply. Ozone concentration in the gas phase was measured following
Table 1 Composition of wheat and rye straw Composition
Wheat, % (w/w)
Rye, % (w/w)
Moisture Cellulose (as glucose) Hemicellulose (as xylose) Acid insoluble lignin Acid soluble lignin
9.0 34.2 20.1 17.1 5.2
5.9 30.9 21.5 22.1 3.2
The experimental work was conducted in three consecutive trials. The first trial was designed to examine the influence of the ozone pretreatment operation factors [Moisture (M), particle size (S), ozone concentration (O), type of biomass (L) and air/O3 flow rate (C)] on straw enzymatic digestibility. Experiments were
Table 2 Experimental levels of the parameters studied in the experimental design for trial s 1 and 3 Factor
Moisture (M) (%w/w) Particle size (S) (mm) Ozone concentration (O) (%w/w) Type of biomass (L) Air/ozone flow rate (C) (L/ha) Basic conditions (B) (% NaOHb) a
Level value Lower (1)
Upper (1)
20 <1 2.7 Wheat 60 0
40 3–5 3 Rye 90 20
Operation variable in trial 1. Operation variable in trial 3. When parameter B is in the upper level, samples were moistened with 20% NaOH solution. The air/ozone flow in this trial experiments was 60 L/h. b
1610
M.T. García-Cubero et al. / Bioresource Technology 100 (2009) 1608–1613
analyzed through a 25–1 factorial design in which the variables were evaluated at two levels (Table 2). Assessing the impact of the main operating factors and their interactions was calculated and laid out in a normal probability plot to draw the most significant factors of the process. Estimating the effect is defined as the change in the response caused by a change in the level of the parameter. The second trial was performed to examine the effect of moisture content during ozone treatment. Water was added at several water-to-straw ratios to obtain materials with different moisture content. Finally, a third trial was conducted to determine the effect of basic conditions on ozonolysis pretreatment by moistening the samples with 20% NaOH. Basic impregnated wheat and rye straw chips were pretreated under eight distinct experimental conditions. These conditions were derived from the same factorial design as in the first trial with a change in parameter air/O3 flow rate (C) for which a value of 60 L/h for pH condition (B) (Table 2) was maintained in this trial.
3. Results and discussion 3.1. Determination of ozonolysis time To set the time required for the reaction, several ozonization tests were run modifying reaction time for rye straw. All these experiments were carried out with an air/ozone flow of 60 L/h, 2.7% w/w ozone, 40% w/w moisture and 3–5 mm particle size. The most striking ozone pretreatment effects observed were acid insoluble and soluble lignin variations. Evolution of these variables was followed to establish reaction time in further experiments. Results obtained showed that ozonization effects remained essentially constant for reaction times above 2.5 h. Acid insoluble lignin (AIL) content decreased from 22.1% (w/w) for non-ozonated material to 12.3% at ozonation times above 2.5 h, whereas acid soluble lignin (ASL) content increased from 3.2% for raw material to 8.1% after 2.5 h. Ozonization pretreatment seems to solubilize more than to degrade the initial lignin present in raw material. From these results a reaction time of 2.5 h was selected for subsequent trials. 3.2. Effect of ozone pretreatment operation variables on straw delignification As explained above, the influence of five parameters on a 25–1 (16 run) factorial design was assessed. This method was used to provide a good insight into pretreatment optimization, establishing which parameters have a significant effect on the ozonization of wheat and rye straw, either directly or by their interactions. Possible inhibitors such as furfural or HMF were not detected in ozonated samples. The impact of ozone on lignin content under the different operating conditions is illustrated in Table 3. All experiments led to a reduction in acid insoluble lignin compared to untreated biomass. ASL content increases with ozone pretreatment compared to both untreated straws, although no clear relationship between ozonization conditions and lignin solubilization was found. Possible structural changes arising from ozone pretreatment were studied analysing the contents of Acid Detergent Fiber (ADF) and Neutral Detergent Fiber (NDF). Results for wheat and rye straw before and after ozonolysis pretreatment of samples under the different operating conditions from the first trial are presented in Fig. 1 as a function of AIL variation in the pretreated samples. The continuous lines in each figure represent the theoretical value that ADF (comprising cellulose and lignin fraction) and NDF (comprising cellulose, hemicellulose and lignin fractions)
Table 3 Results obtained in the 25-1 experimental design for trial 1
msolc MsoLc mSoLc MSolc msoLC MsolC mSolC MSoLC msOLc MsOlc mSOlc MSOLc msOlC MsOLC mSOLC MSO1C
%AIL (%w/w)
%ASL (%w/w)
AIL variation (%w/w)
Effect estimation Parameter
AIL variation
11,8 13,2 16,2 13,5 18,3 12,2 13,5 12,1 17,7 11,8 14,4 11,7 13,9 12,4 17,1 11,2
9,3 9,3 7,2 7,3 6,3 10,4 8,6 7,4 7,7 8,6 8,5 8,5 9,7 8,9 6,6 8,4
32,4 40,3 26,5 22,7 17,1 30,1 22,7 45,4 19,8 32,4 17,7 47,0 20,6 43,8 22,6 35,9
M S c o L CO SO MO OL SC MC CL MS SL ML MSCOL
59,2 2,0 0,6 1,3 24,0 6,2 4,8 19,3 2,5 12,9 13,1 4,4 2,4 18,5 31,3 238,5
M, S, O, L and C are the five considered parameters, e.g. moisture, particle size, ozone concentration, type of biomass and air/ozone flow rate. Upper case letters and lower case letters were respectively used to represent upper and lower level of every considered parameter. %AIL is the acid insoluble lignin obtained for every pretreated samples, %ASL is the acid soluble lignin, AIL variation is the amount of acid insoluble lignin regarding the AIL presented in raw materials and Effect estimation is the change in the response that is produced by a change in the level of the parameter, e.g. acid insoluble lignin variation.
a
wheat straw
100 80
60 % NDF % ADF 40 20 0
b
0
10 20 30 % Acid Insoluble Lignin
40
rye straw
100 80
% NDF 60 % ADF 40 20 0
0
10
20 30 40 %Acid Insoluble Lignin
50
Fig. 1. NDF and ADF as a function of the percentage of Acid Insoluble Lignin removed by the ozonolysis pretreatment of a) wheat straw and b) rye straw. (h)% NDF for wheat straw. (s)% ADF for wheat straw. (e)% NDF for rye straw. (4)% ADF for rye straw. The continuous line indicates the theoretical values if only lignin was degraded. The dotted line shows the linear trend for the experimental points.
should have when eliminating only the lignin fraction calculated from results in Table 3. The dotted row is the line followed by the real points of ADF and NDF obtained in sample analysis. In Fig. 1a), the continuous and dotted lines for ADF wheat straw
1611
M.T. García-Cubero et al. / Bioresource Technology 100 (2009) 1608–1613
overlap, meaning that the theoretical and experimental ADF values are similar. It may thus be inferred from Fig. 1a) that there was no degradation of cellulose due to ozone pretreatment for wheat straw. However, in Fig. 1b) real ADF values for rye straw deviate slightly from the theoretical line, although only in samples with high acid insoluble lignin removal yield. Results thus evidence a small decrease in cellulose (up to 5%) for lignin removals up to 35%, indicating that ozone is not totally selective with lignin. Fig. 1 shows how experimental NDF values deviate from theoretical points for both raw materials. Taking into account that there are no cellulose losses in wheat straw and since the loss of this fraction in rye straw is below 5%, differences between real and theoretical values obtained should be due to hemicellulose solubilization. The higher difference between the dotted and continuous lines corresponds to the samples with a high degree of delignification. Therefore, solubilization of hemicellulose and lignin is likely to occur at the same time. These experimental results were confirmed by chromatographic analysis in some of the pretreated samples (data not shown). This behaviour has also been evidenced by other authors. Binder et al. (1980) and Ghedalia and Miron (1981) observed that losses in cellulose fractions due to pretreatment were below 5% when working at low pH values. Ghedalia and Miron (1981) pointed out that ozone not only led to a decrease in lignin content but also to the solubilization of the hemicellulose fraction converting it into cell solubles. Sun and Cheng (2002) indicated that as lignin content fell, solubilization of hemicellulose increased, likely because both compounds solubilized together in a lignin-hemicellulose complex. From the previous results, acid insoluble lignin concentration was therefore selected as the variable to study the effect of ozone pretreatment on straw digestibility. For each of the 16 runs, the percentage of AIL variation and its main effects and interactions are given. The percentage of AIL variation was calculated taking into account that the percentage of acid insoluble lignin was 17.1 and 22.1 for untreated wheat and rye straw respectively. Statistical treatment of the results in Table 3 revealed that moisture and biomass type (L) were the most significant variables from the studied parameters, always working with an identical operation time (2.5 h). Fig. 2 represents the normal probability plot, in which these effects deviate from the line that goes through the remaining points. The value of the ‘‘biomass type” (L) effect was positive in the normal probability plot, ozonization pretreatment thus removing a higher percentage of acid insoluble lignin in rye than in wheat
straw, although the lowest remaining lignin obtained in the pretreated samples was analogous for both materials (around 11– 12%). A fraction of acid insoluble lignin exposed to ozone can probably be solubilized or degraded, leaving cellulose more accessible to enzymes as suggested by Neely (1984), although there must be another fraction of lignin refractory to ozonization pretreatment. As before, following the factorial design, as the percentage of water content increased, an increase in the removal of acid insoluble lignin for both type of biomass was observed. The remaining parameters studied evidenced no major significance from the point of view of delignification yield and digestibility, although they do need to be taken into account to make this pretreatment feasible from an economical standpoint. Results from the first trial allow us to select standard conditions for the parameters considered to be used in all subsequent experiments, e.g. air/ozone flow rate of 60 L/h, 2.7% w/w ozone concentration and a 3–5 mm particle size. Experiments in the first trial were carried out with two different moisture contents. In order to evaluate the possibility of a non-lineal relationship between delignification and moisture, seven runs with different water content were performed for both straws. Experimental results are shown in Fig. 3. AIL solubilization behaviour during ozonolysis pretreatment as moisture content increases is very similar for both straws. Fig. 3 shows how AIL content diminished as moisture increased up to 30%, likely due to an improvement in mass transfer from air/ozone flow rate to solid surface. An increase in water content above 30% evidenced no significant effects on ozonolysis pretreatment. Working with oak sawdust, Neely (1984) states the optimum range for water content should be 25–35% by weight, and Vidal and Molinier (1988) working with poplar sawdust obtained an optimum water content of 75%, probably due to the higher ozone flow rates employed in the tests. Working with cotton stalks with much higher water content at a solid loading of 10% w/v, Silverstein et al. (2007) obtained no significant changes in lignin, xylan or glucan contents. To study the influence of basic conditions on ozonolysis pretreatment, pH has been included in the factorial design instead of air/O3 flow rate (see Table 2). For statistical interpretation of this influence, the percentage of AIL variation in each sample and the estimation of its principal effects and their interactions are presented in Table 4. Fig. 4 shows the most important factors in ozonolysis pretreatment in basic conditions to be pH and, to a lesser extent, moisture and biomass type. The negative value of parameter B indicates that when basic conditions were used the percentage of acid insoluble lignin removed decreased. Lignin removal ranged between 1.6 and 19.6%
4 3
20
L 1 0 -40
-20
-1
0
20
40
60
80
-2
% Acid Lignin
M
2
15 10 5 0 0
-3 Fig. 2. Normal probability plot of the effects of parameters in the experimental design (trial 1) for AIL variation. X-axis represents the effect estimation and Y-axis the normal probability (depends on the number of experimental runs).
20
40 60 Water Content (%)
80
Fig. 3. Effect of moisture during ozonolysis pretreatment for both straws. Study of the amount of AIL and ASL for each moisture content. (h)% AIL for wheat straw. (s)% ASL for wheat straw. (e)% AIL for rye straw. (4)% ASL for rye straw.
M.T. García-Cubero et al. / Bioresource Technology 100 (2009) 1608–1613
Table 4 Results obtained in the 25–1 experimental design (trial 3)
Msolb MsoLb mSoLb Msolb msoLB MsolB mSolB MSoLB msOLb MsOlb mSOlb MSOLb msOlB MsOLB mSOLB MSO1B
100
%AIL
%ASL
AIL variation
Effect estimation
(%w/w)
(%w/w)
(%w/w)
Parameter
AIL variation
11,8 13,2 16,3 13,5 18,7 17,0 14,8 17,1 17,7 11,8 14,4 11,7 15,9 16,5 19,1 14,1
9,3 9,3 7,2 7,3 5,9 7,1 7,0 6,6 7,7 8,6 8,5 8,5 7,8 8,1 6,6 6,5
32,4 40,3 26,5 22,7 15,6 1,6 15,6 22,6 19,8 32,4 17,7 47,0 9,3 9,7 13,5 19,6
M S B 0 L OB SO MO OL SB MB LB MS SL ML MSOLB
22,8 12,1 65,6 4,1 21,9 0,8 14,6 25,6 10,9 23,1 23,3 6,6 15,7 12,2 21,5 173,2
a B correspond with the parameter ‘‘Basic conditions”. Those experiences which were moistened with 20% NaOH solution.
2
MB
1 0 -90
-40
10
60
-1
B
-2
Fig. 4. Normal probability plot of the effects of the parameters in the experimental design (trial 3) for AIL variation. X-axis represents the effect estimation and Y-axis the normal probability (depends on the number of experimental runs).
for wheat straw and 9.7–22.6% for rye straw, lower than results derived from the first trial. Statistical analysis showed a lower influence of moisture and biomass type, probably due to the significant impact of the pH parameter. Alkaline conditions for ozonolysis pretreatment were used by Binder et al. (1980) in a stirred semibatch reactor. However, the delignification extent obtained diminished noticeably with acid insoluble lignin compositions near to 16– 17% being obtained. Additionally, results from NDF and ADF analysis of ozone pretreated straw in the third trial, when sodium hydroxide was added, showed an increase in cellulose degradation of up to 40% for rye straw and 50% for wheat straw, whereas 50% of the hemicellulose present in both straws was degraded. These results agree with those obtained by Binder et al. (1980), Roncero et al. (2001) and Coca et al. (2005). These authors conclude that under basic conditions hydroxide ions catalyze the decomposition of ozone to yield highly reactive and non-selective hydroxyl radicals, delignification decreases, and degradation of cellulose and hemicellulose increases. 3.3. Enzymatic hydrolysis The yield obtained after enzymatic hydrolysis of the pretreated samples from the first and second trials, defined as the percentage of glucose hydrolyzed, was correlated with the acid insoluble lig-
% Glucose hydrolyzed
1612
80 60 40 20 0
10
15 20 % Acid Insoluble Lignin
25
Fig. 5. Enzymatic hydrolysis yield as a function of the percentage of acid insoluble lignin present in the pretreated samples for wheat straw (h) and rye straw (e).
nin concentration as shown in Fig. 5, for wheat and rye straw respectively, in agreement with results obtained by Zhu et al. (2008) when poplar wood was pretreated and hydrolyzed. The ozonated material from the third trial was not enzymatically hydrolyzed due to the high losses observed in cellulose fraction after ozone pretreatment at high pH. In all cases, when the ozonated materials were hydrolyzed, the yield obtained increased compared to untreated straw. From the two types of straws compared, wheat proved easier to hydrolyze than rye, even when not being treated with ozone. Enzymatic hydrolysis yield improved from 29% corresponding to untreated raw material to 53–88.6% for pretreated wheat straw, corresponding to extreme values of acid insoluble lignin content obtained in ozonated samples, which decreased from 17.1% to 13.5% and 11.2%, respectively. The same behaviour was observed for rye straw. The enzymatic hydrolysis yield improved from 16% corresponding to untreated material to 36–57% for pretreated samples when acid insoluble lignin content was reduced from 22.1% to 18.3% and 12.1%, respectively. Therefore, ozonolysis pretreatment partially removes the acid insoluble lignin content of the biomass, increasing the yield in subsequent enzymatic hydrolysis. The enzymatic hydrolysis yields obtained with ozone pretreated straw in this work are in the same range as the best results found for these raw materials. Sun and Chen (2008) found 90% enzymatic hydrolysis yield when wheat straw was pretreated by aqueous glycerol, with a 70% removal of hemicellulose and 65% of lignin. García-Aparicio et al. (2007), working with steam exploded barley straw (210 °C, 10 min), reported values between 58.2% and 85.7% after hydrolysis times of 24 and 120 h respectively. Rosgaard et al. (2007) obtained lower glucose yields (39% w/w) with pretreated wheat straw using hot water extraction (16% dry matter, 60 °C, 15 min; liquids removed; 180 °C, 10 min; 195 °C, 3 min). In the same work, water impregnated and steam exploded wheat straw (70% dry matter, 1 h, 220 °C and 2.5 min) showed a glucose yield of 30% after 72 h of enzymatic hydrolysis. 4. Conclusions Ozonolysis is an efficient pretreatment for cereal straw. Ozone degrades and/or solubilizes lignin and slightly solubilizes the hemicellulose fraction, improving subsequent enzymatic hydrolysis. Negligible losses of cellulose were detected. Operating in a fixed bed reactor, moisture and biomass type proved the most relevant parameters. Moisture is a reaction controlling parameter for values below 30%. Wheat straw proved easier to hydrolyze than rye, although a similar content of residual lignin after ozone pretreatment was obtained for both essayed materials.
M.T. García-Cubero et al. / Bioresource Technology 100 (2009) 1608–1613
Ozone pretreatment in a basic medium did not prove adequate. Addition of sodium hydroxide reduces the delignification effect and increases degradation of both cellulose and hemicellulose fractions. Acknowledgements The authors wish to thank the Castilla y León Institute of Technological Agriculture (Ref. VA-01/2005), the Regional Ministry of Education at the Castilla y León Regional Government (VA060A06) and the Spanish Ministry of Education and Science (CTQ2006-15217/ PPQ) for financially supporting this research. Irune Indacoechea is also grateful to the Castilla y León Regional Government for providing her Doctorate Scholarship. References Amat, A.M., Arques, A., Miranda, M.A., López, F., 2005. Use of ozone and/or UV in the treatment of effluents from board paper industry. Chemosphere 60, 1111–1117. Binder, A., Pelloni, L., Fiechter, A., 1980. Delignification of straw with ozone to enhance biodegradability. Eur. J. Appl. Microbiol. Biotechnol. 11, 1–5. Coca, M., Peña, M., González, G., 2005. Variables affecting efficiency of molasses fermentation wastewater ozonization. Chemosphere 60, 1408–1415. Contreras, S., 2002. Degradation and biodegradability enhancement of nitrobenzene and 2,4-dichlorophenol by means of advanced oxidation processes based on ozone. PhD Thesis. University of Barcelona. 84-688–2054-7. García-Aparicio, M.P., Ballesteros, M., Manzanares, P., Ballesteros, I., González, A., Negro, M.J., 2007. Xylanase contribution to the efficiency of cellulose enzymatic hydrolysis of barley straw. Appl. Biochem. Biotechnol. 136–140, 353–365. Ghedalia, D.B., Miron, J., 1981. The effect of combined chemical and enzyme treatments on the saccharification and in vitro digestion rate of wheat straw. Biotechnol. Bioeng. 23, 823–831. Goering, H.D., Van Soest, P.J., 1970. Forage fiber analyses. Agric. Handbook., N 379. USDA, Washington DC. Hendriks, A.T.W.M., Zeeman, G., 2008. Pretreatments to enhance the digestibility of lignocellulosic biomass. Bioresour. Technol. doi:10.1016/j.biortech.2008.05.027. Mosier, N., Wyman, C., Dale, B., Elander, R., Lee, Y.Y., Holtzapple, M., Ladish, M., 2005. Features of promising technologies for pretreatment of lignocellulosic biomass. Bioresour. Technol. 96, 673–686. Mussato, S.I., Fernandes, M., Milagres, A.M.F., Roberto, I.C., 2008. Effect of hemicellulose and lignin on enzymatic hydrolysis of cellulose from brewer‘s spent grain. Enzyme Microb. Technol. 34, 124–129.
1613
National Renewable Energy Laboratory (NREL), 1995. Chemical Analysis and Testing Laboratory Analytical Procedures: LAP-002, LAP-003, LAP-004. NREL, Golden, CO, USA.