- Email: [email protected]
Production of xylose from oil palm empty fruit bunch fiber using sulfuric acid
Biochemical Engineering Journal 30 (2006) 97–103
Production of xylose from oil palm empty fruit bunch fiber using sulfuric acid S.H.A. Rahman, J.P. Choudhury ∗ , A.L. Ahmad School of Chemical Engineering, Engineering Campus, Universiti Sains Malaysia, 14300 Nibong Tebal, SPS Penang, Malaysia Received 30 May 2005; received in revised form 20 January 2006; accepted 18 February 2006
Abstract Oil palm empty fruit bunch fiber is a lignocellulosic waste from palm oil mills. It is a potential source of xylose which can be used as a raw material for production of xylitol, a high value product. The increasing interest on use of lignocellulosic waste for bioconversion to fuels and chemicals is justifiable as these materials are low cost, renewable and widespread sources of sugars. Batch hydrolysis of oil palm empty fruit bunch fiber was performed at operating temperature 120 ◦ C using various concentration of sulfuric acid (2–6%) and reaction time (0–90 min). Concentration of xylose, glucose, furfural and acetic acid in the resulting hydrolysate were determined. Kinetic parameters of mathematical models were obtained in order to predict concentration of xylose, glucose, furfural, acetic acid in the hydrolysate and to optimize the process. Optimum H2 SO4 concentration and reaction time obtained under operating temperature of 120 ◦ C was 6% and 15 min, respectively. Optimum concentration of xylose, glucose, furfural and acetic acid found in the hydrolysate were 29.4, 2.34, 0.87 and 1.25 (g/l), respectively. © 2006 Elsevier B.V. All rights reserved. Keywords: Oil palm empty fruit bunch; Hydrolysis; Xylose; Kinetic parameters; Optimisation; Cellulose; Glucose
1. Introduction Bioconversion of lignocellulosic waste materials to chemicals and fuels are receiving interest as they are low cost, renewable and widespread in nature [1]. Malaysia is well known for its potential in renewable resources such as oil palm waste, sugar cane bagasse and rice straw. At present Malaysia is the largest exporter of palm oil in the international market. In the process of extraction of palm oil from oil palm fruit, a lignocellulosic material oil palm empty fruit bunch (OPEFB) is generated as a waste product. Approximately 15 million tons of OPEFB biomass waste is generated annually throughout Malaysia by palm oil mills. In practice this biomass is burned in incinerators by palm oil mills which not only creates environmental pollution problems in nearby localities but also it offers limited value to the industry. The OPEFB biomass contains cellulose, hemicellulose and lignin. It is estimated that OPEFB biomass is comprised of 24% xylan, a sugar polymer made of pentose sugar xylose. This xylose can be used as substrate for production of a wide variety of compounds by chemical and biochemical processes
∗
Corresponding author. Tel.: +60 4 5937788x6424; fax: +60 4 5941013. E-mail address: jyoti [email protected] (J.P. Choudhury).
1369-703X/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.bej.2006.02.009
[2,3]. One such compound is xylitol, which is extensively used in food, pharmaceutical and thin coating applications [4]. The most important application of xylitol is its use as an alternative sweetener in foods for diabetic patients [5]. Other important uses of xylitol are: as an anticariogenic agent in tooth paste formulations; as thin coatings on chewing vitamin tablets; in mouth washes, beverages and in bakery products [6,7]. Research investigations on dilute acid hydrolysis of various raw materials such as sugar cane bagasse, sorghum straw, corncobs and eucalyptus wood have been carried out by several researchers [8–11]. From the research studies it was revealed that under controlled treatment conditions, acid hydrolysis of lignocellulosic biomass mainly produced xylose from xylan with cellulosic and lignin fractions remaining unaltered. The solid residue can further be utilized for production of ethanol or in pulp processing for making high grade paper. In the hydrolysis process it is understood that initially the lignin protective layer around the hemicellulose fiber is softened under elevated temperature and pressure which allows the acid to penetrate the layer and hydrolyze the amorphous xylan to form xylose. On the other hand the condition is not severe enough to hydrolyze the crystalline structure of cellulose which remains as insoluble solid. Although xylose was the main sugar obtained from hemicellulose, other byproducts such as glucose, acetic acid, furfural, etc.
98
S.H.A. Rahman et al. / Biochemical Engineering Journal 30 (2006) 97–103
were also produced in low amount during the hydrolysis process [10,12]. It was also reported that amount of sugar released during the hydrolysis process, depended on type of raw material and operation conditions of the experiment. Acid concentration is an important parameter for release of sugars whereas temperature is mainly responsible for decomposition of sugars to various byproducts [13]. The use of hydrolysate as a source of xylose sugar in fermentation process depends on types of byproducts present in hydrolysate. It is known that acetic acid and furfural are potential inhibitors to yeast metabolism. When these compounds are present in the hydrolysate, they inhibit the fermentation process by causing cell morphological change or ultimate death of the organism [14]. To keep the concentration of byproducts (glucose, acetic acid and furfural) in the hydrolysate at low level it is necessary to run the hydrolysis reaction at less severe conditions. To our knowledge, there are no reports on acid hydrolysis of OPEFB biomass. The objective of the present investigation was to study the hydrolysis of OPEFB biomass with dilute sulfuric acid in the range of 2–6% at operating temperature 120 ◦ C and to determine the kinetic parameters in order to predict the release of xylose, glucose, acetic acid and furfural in the resulting hydrolysate and to optimize the process. Operating temperature (120 ◦ C) and acid concentration range (2–6%) were selected on the basis of previous research reports on hydrolysis of various lignocellulosic raw materials [2,4,11,15–17]. 2. Materials and methods 2.1. Raw material Oil palm empty fruit bunch (OPEFB) fiber was collected from local palm oil mill (United Oil Palm Industries Sdn Bhd, Malaysia), sun dried and ground to a particle size <1 mm. The homogenized OPEFB biomass was then oven dried at 105 ◦ C for overnight and was analyzed following standard method for determination of its main composition [18].
Chromatograph (GLC, Perkin-Elmer) using 80/120 carbopack column with nitrogen as carrier gas and detection was done by FID. The column and injector temperatures were maintained at 225 and 250 ◦ C, respectively while oven was operated at 175 ◦ C. All experiments were carried out in triplicate and data were expressed in average values. Experimental data were fitted to the proposed equations and non-linear regression analyses were performed using Newton’s method (Solver, Microsoft Excel 2000, Microsoft Corporation, Redmont, WA, USA). 2.4. Kinetic models Various kinetic models for acid hydrolysis of lignocellulose are available in literature. The model proposed on hydrolysis of cellulose involves polymer glucan of cellulose is degraded to monomer glucose which is subsequently converted to decomposition products [19,20]. The process can be represented as: glucan(s) → glucose → decomposition products The acid hydrolysis of hemicellulose proposed by different authors were, originated from two different kinetic mechanisms [16,17,19–21]. In one mechanism it is assumed that xylan is first converted to xylooligosaccharide, which in turn gives xylose on further hydrolysis and xylose is subsequently decomposed to furfural. Other mechanism does not include the intermediate formation of xylooligosaccharide, where xylan is converted to xylose which in turn is dehydrated to furfural. The accuracy of final results for both the mechanisms is same. The simplest kinetic model for hemicellulose hydrolysis involves pseudohomogeneous irreversible first-order series reactions, which is represented by the following scheme: xylan(s) → xylose(aq) → decomposition products Thus it can be generalized as k1
k2
polymer−→monomer−→decomposition products 2.2. Acid hydrolysis Acid hydrolysis of OPEFB biomass was carried out in 125 ml Erlenmyer flasks. The media consisted of 2–6 g H2 SO4 /100 g liquor using a charge of 1 g OPEFB fiber/8 g liquor on dry basis. Operating temperature of hydrolysis was 120 ◦ C and samples were collected at various time intervals in the range of 0–90 min. The samples were diluted with water and insoluble solids were separated from aqueous solution by filtration. The filtrate was analyzed for xylose, glucose, acetic acid and furfural. 2.3. Analytical methods Xylose and glucose in the acid hydrolysate were analyzed by High Performance Liquid Chromatograph (HPLC, SHIMADZU) using SUPELCOSIL LC-NH2 column and RI detector. Aqueous acetonitrile (75%) was used as mobile phase with flow rate of 1.5 ml min−1 and oven temperature was maintained at 50 ◦ C. Acetic acid and furfural were analyzed by Gas Liquid
(1)
where k1 is the rate of monomer released (min−1 ) and k2 is the rate of monomer decomposed (min−1 ). Based on this reaction model and solving differential equations, monomer concentration (M) as a function of time (t) can be represented by M = [k1 P0 /(k2 − k1 )](e−k1 t − e−k2 t ) + M0 e−k2 t
(2)
where P and M represents concentration of polymer and monomer, respectively. The subscript 0 represents at time t = 0. Assuming M0 to be nearly equal to 0, Eq. (2) can be simplified as M = [k1 P0 /(k2 − k1 )](e−k1 t − e−k2 t )
(3)
The kinetic model for cellulose hydrolysis involves pseudohomogeneous irreversible first order series reactions, which can be represented by the following scheme: k3
k4
glucan(s)−→glucose(aq)−→decomposition products
(4)
S.H.A. Rahman et al. / Biochemical Engineering Journal 30 (2006) 97–103
99
As the operating conditions in the present investigation were not favorable for decomposition of glucose, hence the model can be modified as k3
glucan(s)−→glucose(aq)
(5)
Solving differential equations, concentration of glucose (G) as a function of time can be represented as G = G0 (1 − e−k3 t )
(6)
where k3 is the glucose generation rate (min−1 ) and G0 is the potential glucose concentration which can be estimated by regression analysis. The model for furfural concentration (F) as a function of time can be represented as F = F0 (1 − e−k4 t )
(7)
where F0 is the potential concentration of furfural which can be determined by regression analysis and k4 is the furfural generation rate (min−1 ). Acetyl groups originated from hemicellulose and model for acetic acid generation can be represented by the following scheme: k5
acetyl groups−→acetic acid
(8)
Concentration of acetic acid (A) in the hydrolysate as a function of time can be represented as follows: A = A0 (1 − e−k5 t )
(9)
where A0 is potential concentration of acetic acid which can be obtained by regression and k5 is the acetic acid generation rate (min−1 ). Eqs. (3), (6), (7) and (9) were applied to sulfuric acid hydrolysis of OPEFB fiber. Kinetic parameters and constants were evaluated statistically and determination coefficient R2 of each was obtained in order to find out the reliability of the models. 3. Results and discussion
Fig. 1. Effect of sulfuric acid concentration and reaction time on experimental and predicted concentrations of xylose released for acid hydrolysis of OPEFB fiber at 120 ◦ C.
where Xp0 is the initial amount of xylan polymer present in the OPEFB fiber on dry basis (24.01 g xylan/100 g OPEFB fiber), 150/132 is the stoichiometric factor and LSR is liquid solid ratio (8 g liquid/g of OPEFB fiber). Analysis of acid hydrolysate showed that highest release of xylose from hemicellulose was 31.1 (g/l) when acid concentration was 2% which was 91.2% of potential concentration of xylose. On the other hand with 4 and 6% sulfuric acid, maximum xylose released were 30.7 and 30.27 (g/l), respectively. This is shown in Fig. 1. From the figure it was observed that with increase in acid concentration, xylose concentration attained a peak value for any acid concentration studied and with prolong reaction time, concentration of xylose in the resulting hydrolysate was ultimately decreased. Experimental results suggest that probably there exist some decomposition reaction which leads to dehydration of xylose to furfural. During acid hydrolysis other sugar glucose was also released but in low concentration which is shown in Fig. 2. The highest amount of glucose released in the hydrolysate was 4.1 (g/l) when acid concentration was 6%. On the other hand with 2 and
Analysis of the empty fruit bunch fiber was carried out for determination of principal components using quantitative acid hydrolysis following standard method [18] which is shown in Table 1, expressed on an oven-dry basis. If we assume that polymer xylan is completely converted to xylose without formation of any decomposition products, then P0 can be represented by equivalent amount of xylose by following calculation: P0 = [Xp0 × 150 × 10]/[132 × LSR] = 34.1 g xylose/l (10) Table 1 Main components of oil palm empty fruit bunch fiber (on an oven-dry basis) Main fraction
Composition (%)
Glucan Xylan Lignin (acid insoluble) Ash Others
42.85 24.01 11.70 0.52 20.92
Fig. 2. Dependence of experimental and predicted concentrations of glucose released on acid concentration and reaction time for sulfuric acid hydrolysis of OPEFB fiber at 120 ◦ C.
100
S.H.A. Rahman et al. / Biochemical Engineering Journal 30 (2006) 97–103 Table 2 Kinetics and statistical parameters of components released during H2 SO4 hydrolysis of OPEFB fiber at 120 ◦ C
Fig. 3. Effect of sulfuric acid concentration and reaction time on experimental and predicted concentrations of furfural released for acid hydrolysis of OPEFB fiber at 120 ◦ C.
4% H2 SO4 , maximum glucose released were 1.8 and 3.2 (g/l), respectively. The release of glucose could be either from hemicellulose or cellulose chain. Hence equivalent concentration of glucose from glucan could not be determined. The amount of furfural generated as decomposition product in the resulting hydrolysate is shown in Fig. 3. From the figure it was evident that when H2 SO4 concentration was increased from 2 to 6%, furfural concentration was increased in the resulting hydrolysate. The highest concentration of furfural noticed was 3.3 (g/l) when acid concentration was 6%. This result was in agreement with the reduction of xylose concentration at higher acid concentration which was observed in Fig. 1. It is known that during hydrolysis process acetic acid is generated in the hydrolysate from acetyl groups of hemicellulose. The maximum and minimum concentration of acetic acid in the resulting hydrolysate was found to be 3.93 and 0.59 (g/l) when sulfuric acid concentrations were 6 and 2%, respectively which is shown in Fig. 4.
2% H2 SO4
4% H2 SO4
6% H2 SO4
Xylose k1 (min−1 ) k2 (min−1 ) R2
0.0451 0.0015 0.99
0.0827 0.0029 0.99
0.1695 0.0057 0.99
Glucose k3 (min−1 ) G0 R2
0.0238 3.50 0.99
0.0332 3.70 0.99
0.0518 4.43 0.99
Furfural k4 (min−1 ) F0 R2
0.0031 4.88 0.98
0.0063 6.57 0.99
0.0118 5.28 0.98
Acetic acid k5 (min−1 ) A0 R2
0.0064 6.1 0.96
0.0142 5.55 0.98
0.0189 5.02 0.99
3.1. Kinetic model of xylose concentration Kinetic and statistical parameters obtained on hydrolysis of OPEFB fiber at 120 ◦ C with sulfuric acid in the concentration range of 2–6% is shown in Table 2. Experimental and predicted data for xylose release with various acid concentration and reaction time is shown in Fig. 1. From the table it was evident that with sulfuric acid concentration of 2% xylose generation rate k1 was higher than the decomposition rate k2 which are 0.0451 and 0.0015 (min−1 ), respectively. The determination coefficient R2 showed a good agreement between experimental and predicted data for all regressions. It was also evident that with increase in acid concentration values of k1 and k2 were also increased, which indicated that with increase in acid concentration a shorter reaction time was needed to obtain maximum release of xylose in the resulting hydrolysate. A generalized model for predicting all products was developed by relating kinetic parameters with sulfuric acid concentration by empirical equation kg = k0 Can
Fig. 4. Dependence of experimental and predicted concentrations of acetic acid released on sulfuric acid concentration and reaction time for acid hydrolysis of OPEFB fiber at 120 ◦ C.
(11)
where g is in the range of 1–5; k0 and n the regression parameters; Ca is H2 SO4 concentration expressed in % w/w. Thus generalized model for xylose release is represented by empirical Eq. (12) where k1 is correlated with acid concentration Ca . Similarly xylose decomposition rate k2 is represented by empirical Eq. (13). The determination coefficients R2 for both the parameters were in good agreement which is shown in Table 3. By combining Eqs. (12) and (13) with the models of xylose formation and degradation it is possible to predict xylose concentration at any time and acid concentration within the time period (0–90 min) and acid concentration (2–6%) studied. The generalized model predicted that maximum xylose concentration of more than 29 (g/l) could be obtained with 2, 4 and 6% sulfuric acid at reaction temperature of 120 ◦ C when hydrolysis reaction time were 90, 45 and 15 min, respectively. The dependence of xylose concentration with acid concentration and reaction time is shown
S.H.A. Rahman et al. / Biochemical Engineering Journal 30 (2006) 97–103
101
Table 3 Generalized models for prediction of kinetic parameters of OPEFB fiber hydrolysis with H2 SO4 at 120 ◦ C Products
Models
R2
Xylose
k1 k2 k3 k4 k5
0.97 0.98 0.95 0.98 0.98
Glucose Furfural Acetic acid
= 0.013 Ca1.4217 (12) = 0.0005 Ca1.3849 (13) = 0.0127 Ca0.768 (14) = 0.0011 Ca1.3388 (15) = 0.0037 Ca0.9187 (16)
by response surface in Fig. 5. Thus from the response surface it is possible to select shorter reaction time to obtain maximum release of xylose with minimum decomposition products in the resulting hydrolysate. Again by analyzing different response surfaces it is possible to reach optimum condition of the hydrolysis process. 3.2. Kinetic model of glucose concentration In the model of glucose release (Eq. (6)), G0 could not be obtained experimentally as glucose could have released from both hemicellulose and cellulose. Hence G0 and kinetic parameter k3 were obtained by regression and determination coefficient R2 showed was in good agreement with experimental and predicted data. The values of G0 obtained were in the range of 3.5–4.43 (g/l). The values of k3 were found to be in the range of 0.0238–0.0518 (min−1 ), which is shown in Table 2. It was observed that the values of k3 were increased with increase in acid concentration. Hence a generalized model for prediction of glucose concentration was developed which was correlated with acid concentration and kinetic parameter k3 . This is presented by empirical Eq. (14) and determinant coefficient R2 was well fitted which is given in Table 3. From the table it was observed that the value of regression parameter n for glucose generation
Fig. 6. Effect of sulfuric acid concentration and reaction time on generalized model for prediction of glucose concentration.
rate k3 was 0.768 which was lower than that of corresponding value of regression parameter (n = 1.4217) for xylose generation k1 . It is known that hemicellulose is amorphous in character whereas cellulose is crystalline in nature. Moreover hydrolysis of cellulose is strongly dependent on degree of crystallinity and swelling state of cellulose [22]. The results suggested that probably heterogeneous reaction occurred during release of glucose from glucan. Hence the effect of acid on glucose generation rate k3 was not similar in nature compared to that of effect of acid on xylose generation rate k1 . Experimental and predicted data of glucose generation in the resulting hydrolysate is shown in Fig. 2. Thus with the help of Eq. (14) and glucose generation model, it is possible to predict glucose concentration within the experimental range studied. The models predicted highest glucose concentration of 4.38 (g/l) at most severe reaction conditions (6% acid concentration and 90 min reaction time) and greater than 1.0 (g/l) could be obtained if the reaction was carried out for time period (t < 30 min) for any acid concentration studied. The response surface of the generalized model for glucose generation with increase in acid concentration and reaction time is shown in Fig. 6. From the response surface it was observed that there was no decomposition reaction occurred during the hydrolysis process. To obtain maximum xylose concentration in the resulting hydrolysate, it is necessary to keep glucose concentration as low as possible. Hence response surface in this respect can help to conduct the experiment in the range (shorter reaction time) where glucose concentration can be kept at low level for maximizing xylose concentration in the resulting hydrolysate. 3.3. Kinetic model of furfural concentration
Fig. 5. Effect of sulfuric acid concentration and reaction time on generalized model for prediction of xylose concentration.
In the acid hydrolysis of OPEFB biomass furfural is the principal degradation product of xylose. Kinetic and statistical
102
S.H.A. Rahman et al. / Biochemical Engineering Journal 30 (2006) 97–103
Fig. 7. Effect of sulfuric acid concentration and reaction time on generalized model for prediction of furfural concentration.
parameters obtained for furfural is shown in Table 2. It was found that values of F0 and kinetic parameter k4 were within the range of 4.88–6.57 (g/l) and 0.0034 to 0.0118 (min−1 ), respectively. The determinant coefficients R2 were well fitted with furfural generation model. Experimental and predicted data is shown in Fig. 3. It was noted that furfural formation rate k4 was higher than xylose degradation rate. It was also noted that kinetic parameter k4 was increased with increase in acid concentration. A generalized model was developed to correlate kinetic parameter k4 with acid concentration for prediction of furfural concentration at any time and acid concentration within the operating range studied. The empirical Eq. (15) represents the generalized model which is shown in Table 3. The graph of generalized model for furfural formation is shown in Fig. 7. From the response surface it was observed that with increase in reaction time and acid concentration, furfural concentration in the resulting hydrolysate was increased. It is known that furfural is an inhibitory compound to fermentation process and hence its concentration in the hydrolysate should be minimized to facilitate optimum use of hydrolysate for xylitol production. From the response surface it can be interpreted that it is better to conduct hydrolysis process with higher acid concentration and lower reaction time to minimize the formation of furfural in the resulting hydrolysate. In this case generalized model can help to predict concentration of furfural and hence selection of acid concentration and reaction time can be done in order to obtain maximum concentration of xylose in the hydrolysate while keeping furfural concentration at minimum level. 3.4. Kinetic model of acetic acid concentration In acid hydrolysis of OPEFB fiber, acetic acid is mainly released from acetyl groups of hemicellulose. Kinetic and statistical parameters obtained for acetic acid generation is shown in Table 2. Experimental and predicted data were well fitted
Fig. 8. Effect of sulfuric acid concentration and reaction time on generalized model for prediction of acetic acid concentration.
with the model as is evident from R2 value. Experimental and predicted data for the acid hydrolysis of OPEFB fiber at reaction temperature of 120 ◦ C is shown in Fig. 4. It was observed that value of A0 and k5 were within the range of 6.1–5.02 and 0.0064–0.0189 (min−1 ), respectively. It was also observed that the value of kinetic parameter k5 was increased with increase in acid concentration. A generalized model was developed to correlate k5 with acid concentration for prediction of acetic acid concentration within the experimental range of reaction time (0–90 min) and acid concentration (2–6%) studied. This is presented by empirical Eq. (16) as shown in Table 3. The value of regression parameter n for acetic acid generation k5 obtained was 0.9187 which was lower than the value of regression parameter (n = 1.4217) for xylose generation rate k1 . It is reported that effect of acid on acetyl removal from hemicellulose is lower compared to that of effect of acid on xylan removal which may be the reason for obtaining different n values [23]. The response surface of generalized model for acetic acid released in the hydrolysate is shown in Fig. 8. From the figure it can be interpreted that acetic acid concentration was increased with increase in sulfuric acid concentration and reaction time. Hence to maximize xylose concentration in the resulting hydrolysate it is better to conduct the experiment at high sulfuric acid concentration and low reaction time so that concentration of acetic acid in the hydrolysate can be kept at low level. The generalized models of xylose, glucose, furfural and acetic acid on dilute sulfuric acid hydrolysis of OPEFB fiber predicted several options for obtaining high xylose concentration in the resulting hydrolysate. Thus when the hydrolysis was conducted with 2% sulfuric acid at reaction temperature of 120 ◦ C for 90 min reaction, then xylose, glucose, furfural and acetic acid predicted in the resulting hydrolysate were 29.96, 2.99, 1.08 and 2.82 (g/l), respectively. Whereas when the experiment was conducted with 4% acid for 45 min reaction time the concentration of xylose, glucose, furfural and acetic acid predicted
S.H.A. Rahman et al. / Biochemical Engineering Journal 30 (2006) 97–103
in the resulting hydrolysate were 29.83, 2.99, 1.78 and 2.49 (g/l), respectively. On the other hand when the reaction was carried out with 6% acid for 15 min of reaction time the corresponding concentrations of xylose, glucose, furfural and acetic acid in the acid hydrolysate predicted were 29.4, 2.34, 0.87 and 1.25 (g/l), respectively. Hence the optimum operating condition for production of xylose from OPEFB fiber is to conduct the hydrolysis process at 120 ◦ C with 6% sulfuric acid for 15 min reaction time where all the byproducts can be controlled to minimum level. 4. Conclusion Oil palm empty fruit bunch fiber is a palm oil mill waste. It contains 24% xylan, which is a promising source of xylose. Acid hydrolysis of OPEFB fiber was conducted at operating temperature 120 ◦ C under various H2 SO4 acid concentration (2–6%) and reaction time (0–90 min). The maximum concentration of xylose, glucose, furfural and acetic acid obtained in the resulting hydrolysate were 31.1, 4.1, 3.3 and 3.93 (g/l), respectively. Kinetic parameters of mathematical models were determined in order to predict concentration of xylose, glucose, furfural and acetic acid in the resulting hydrolysate and to optimize the hydrolysis process. The optimum acid concentration and reaction time obtained under operating temperature 120 ◦ C were 6% and 15 min, respectively. The optimum concentration of xylose, glucose, furfural and acetic acid released in the hydrolysate were 29.4, 2.34, 0.87 and 1.25 (g/l), respectively. The product Xylose can further be converted to xylitol by bioconversion process. Thus with optimum use of OPEFB fiber, it will not only solve the environmental pollution problem but also will give back a high value product to the palmoil industry. Acknowledgement The authors are grateful to the Ministry of Science, Technology and Environment, Malaysia Government for the financial support of this work under long term IRPA grant (Project: 0302-05-2215 EA008). References [1] R.C. Kuhad, A. Singh, Current and future prospects, Crit. Rev. Biotech. 13 (1993) 151–172. [2] C.E. Wyman, Ethanol from lignocellulosic biomass: Technology, economics and opportunities, Bioresour. Technol. 50 (1994) 3–16. [3] J.B.A. Silva, I.M. deMancilha, M.C.D. Vannetti, M.A. Teixeira, Microbial protein production by Paecilomyces variotii cultivated in eucalyptus hemicellulosic hydrolysate, Bioresour. Technol. 52 (1995) 197–200.
103
[4] J.C. Parajo, H. Dominguez, J.M. Dominguez, Production of xylitol from raw wood hydrolysate by Debaryomyces hansenii NRRL Y-7426, Bioprocess Eng. 13 (1995) 125–131. [5] T. Pepper, P.M. Olinger, Xylitol in sugar-free confections, Food Technol. 10 (1988) 98–106. [6] K.K. Makinen, Dietary prevention of dental caries by xylitol-clinical effectiveness and safety, J. Appl. Nutr. 44 (1992) 16–28. [7] L. Hyvonen, P. Koivstoinen, Food technological evaluation of xylitol, Adv. Food Res. 28 (1982) 373–403. [8] I.C. Roberto, M.G.A. Felipe, I.M. Mancilha, M. Vitolo, S. Sato, S.S. Silva, Xylitol production by Candida guilliermondii as an approach for the utilization of agro industrial residues, Bioresour. Technol. 51 (1995) 255–257. [9] J.M. Dom´ınguez, N. Cao, C.S. Gong, G.T. Tsao, Dilute acid hemicellulose hydrolysates from corn cobs for xylitol production by yeast, Bioresour. Technol. 61 (1997) 85–90. [10] G. Garrote, H. Dominguez, J.C. Parajo, Generation of xylose solutions from eucalyptus wood by autohydrolysis-posthydrolysis processes: posthydrolysis kinetics, Bioresour. Technol. (2001) 155–164. [11] B.P. Lavarack, G.J. Griffin, D. Rodman, The acid hydrolysis of sugarcane bagasse hemicellulose to produce xylose, arabinose, glucose and other products, Biomass Bioenerg. 23 (2002) 367–380. [12] S.S. Silva, M.G.A. Felipe, J.B.A. Silva, A.M.R. Prata, Acid hydrolysis of eucalyptus grandis chips for microbial production of xylitol, Process Biochem. 33 (1998) 63–67. [13] M. Neureiter, H. Danner, C. Thomasser, B. Saidi, R. Braun, Dilute acid hydrolysis of sugar cane bagasse at varying conditions, Appl. Biochem. Biotechnol. 98 (2002) 49–58. [14] C. Van Zyl, B.A. Prior, J.C. Du Preez, Acetic acid inhibition of dxylose fermentation by Pichia stripitis, Enzyme Microb. Technol. 13 (1991) 82–86. [15] J.C. Parajo, V. Santos, F. del Rio, Hydrolysis of the cellulosic fraction of pinus pinaster wood. II. Production at higher pressure than atmospheric one, Afinidad 52 (1995) 267–274. [16] G. Marton, J. Dencs, L. Szokonya, Principles of Bio Mass Refining: in Handbook of Heat and Mass Transfer, Gulf Publishing, Houston, 1989, 609–652. [17] S.J. Tellez-Luis, J.A. Ramirez, M. Vazquez, Mathematical modelling of hemicellulose sugar production from sorghum straw, Jl. Food Engg. 52 (2002) 285–291. [18] Chemical Analysis and Testing Standard Procedure, National Renewable Energy Laboratories, Golden, Co., NREL (1996), No. 002-004. [19] J.F. Saeman, Kinetics of wood saccharification, hydrolysis of cellulose and decomposition of sugars in dilute acid at high temperature, Ind. Eng. Chem. 37 (1945) 43–52. [20] D.F. Root, J.F. Saeman, J.F. Harris, Chemical conversion of wood residues. Part II. Kinetics of acid-catalyzed conversion of xylose to furfural, Forest Prod. J. 9 (1959) 158–165. [21] F. Carrasco, C. Roy, Kinetic study of dilute-acid-pre-hydrolysis of xylan containing biomass, Wood Sci. Technol. 26 (1992) 189–208. [22] Q. Xiang, Y.Y. Lee, P.O. Pettersson, R.W. Torget, Heterogeneous aspects of acid hydrolysis of ␣-cellulose, Appl. Biochem. Biotechnol. 105–108 (2003) 505–514. [23] J.F. Harris, A.J. Baker, A.H. Connor, T.W. Jeffries, J.L. Minor, R.C. Petersen, R.W. Scott, E.L. Springer, T.H. Wegner, J.I. Zerbe, General technical report FPL-45, US Department of Agriculture Forest Products Laboratory, Madison, WI, 1985.
Production of xylose from oil palm empty fruit bunch fiber using sulfuric acid S.H.A. Rahman, J.P. Choudhury ∗ , A.L. Ahmad School of Chemical Engineering, Engineering Campus, Universiti Sains Malaysia, 14300 Nibong Tebal, SPS Penang, Malaysia Received 30 May 2005; received in revised form 20 January 2006; accepted 18 February 2006
Abstract Oil palm empty fruit bunch fiber is a lignocellulosic waste from palm oil mills. It is a potential source of xylose which can be used as a raw material for production of xylitol, a high value product. The increasing interest on use of lignocellulosic waste for bioconversion to fuels and chemicals is justifiable as these materials are low cost, renewable and widespread sources of sugars. Batch hydrolysis of oil palm empty fruit bunch fiber was performed at operating temperature 120 ◦ C using various concentration of sulfuric acid (2–6%) and reaction time (0–90 min). Concentration of xylose, glucose, furfural and acetic acid in the resulting hydrolysate were determined. Kinetic parameters of mathematical models were obtained in order to predict concentration of xylose, glucose, furfural, acetic acid in the hydrolysate and to optimize the process. Optimum H2 SO4 concentration and reaction time obtained under operating temperature of 120 ◦ C was 6% and 15 min, respectively. Optimum concentration of xylose, glucose, furfural and acetic acid found in the hydrolysate were 29.4, 2.34, 0.87 and 1.25 (g/l), respectively. © 2006 Elsevier B.V. All rights reserved. Keywords: Oil palm empty fruit bunch; Hydrolysis; Xylose; Kinetic parameters; Optimisation; Cellulose; Glucose
1. Introduction Bioconversion of lignocellulosic waste materials to chemicals and fuels are receiving interest as they are low cost, renewable and widespread in nature [1]. Malaysia is well known for its potential in renewable resources such as oil palm waste, sugar cane bagasse and rice straw. At present Malaysia is the largest exporter of palm oil in the international market. In the process of extraction of palm oil from oil palm fruit, a lignocellulosic material oil palm empty fruit bunch (OPEFB) is generated as a waste product. Approximately 15 million tons of OPEFB biomass waste is generated annually throughout Malaysia by palm oil mills. In practice this biomass is burned in incinerators by palm oil mills which not only creates environmental pollution problems in nearby localities but also it offers limited value to the industry. The OPEFB biomass contains cellulose, hemicellulose and lignin. It is estimated that OPEFB biomass is comprised of 24% xylan, a sugar polymer made of pentose sugar xylose. This xylose can be used as substrate for production of a wide variety of compounds by chemical and biochemical processes
∗
Corresponding author. Tel.: +60 4 5937788x6424; fax: +60 4 5941013. E-mail address: jyoti [email protected] (J.P. Choudhury).
1369-703X/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.bej.2006.02.009
[2,3]. One such compound is xylitol, which is extensively used in food, pharmaceutical and thin coating applications [4]. The most important application of xylitol is its use as an alternative sweetener in foods for diabetic patients [5]. Other important uses of xylitol are: as an anticariogenic agent in tooth paste formulations; as thin coatings on chewing vitamin tablets; in mouth washes, beverages and in bakery products [6,7]. Research investigations on dilute acid hydrolysis of various raw materials such as sugar cane bagasse, sorghum straw, corncobs and eucalyptus wood have been carried out by several researchers [8–11]. From the research studies it was revealed that under controlled treatment conditions, acid hydrolysis of lignocellulosic biomass mainly produced xylose from xylan with cellulosic and lignin fractions remaining unaltered. The solid residue can further be utilized for production of ethanol or in pulp processing for making high grade paper. In the hydrolysis process it is understood that initially the lignin protective layer around the hemicellulose fiber is softened under elevated temperature and pressure which allows the acid to penetrate the layer and hydrolyze the amorphous xylan to form xylose. On the other hand the condition is not severe enough to hydrolyze the crystalline structure of cellulose which remains as insoluble solid. Although xylose was the main sugar obtained from hemicellulose, other byproducts such as glucose, acetic acid, furfural, etc.
98
S.H.A. Rahman et al. / Biochemical Engineering Journal 30 (2006) 97–103
were also produced in low amount during the hydrolysis process [10,12]. It was also reported that amount of sugar released during the hydrolysis process, depended on type of raw material and operation conditions of the experiment. Acid concentration is an important parameter for release of sugars whereas temperature is mainly responsible for decomposition of sugars to various byproducts [13]. The use of hydrolysate as a source of xylose sugar in fermentation process depends on types of byproducts present in hydrolysate. It is known that acetic acid and furfural are potential inhibitors to yeast metabolism. When these compounds are present in the hydrolysate, they inhibit the fermentation process by causing cell morphological change or ultimate death of the organism [14]. To keep the concentration of byproducts (glucose, acetic acid and furfural) in the hydrolysate at low level it is necessary to run the hydrolysis reaction at less severe conditions. To our knowledge, there are no reports on acid hydrolysis of OPEFB biomass. The objective of the present investigation was to study the hydrolysis of OPEFB biomass with dilute sulfuric acid in the range of 2–6% at operating temperature 120 ◦ C and to determine the kinetic parameters in order to predict the release of xylose, glucose, acetic acid and furfural in the resulting hydrolysate and to optimize the process. Operating temperature (120 ◦ C) and acid concentration range (2–6%) were selected on the basis of previous research reports on hydrolysis of various lignocellulosic raw materials [2,4,11,15–17]. 2. Materials and methods 2.1. Raw material Oil palm empty fruit bunch (OPEFB) fiber was collected from local palm oil mill (United Oil Palm Industries Sdn Bhd, Malaysia), sun dried and ground to a particle size <1 mm. The homogenized OPEFB biomass was then oven dried at 105 ◦ C for overnight and was analyzed following standard method for determination of its main composition [18].
Chromatograph (GLC, Perkin-Elmer) using 80/120 carbopack column with nitrogen as carrier gas and detection was done by FID. The column and injector temperatures were maintained at 225 and 250 ◦ C, respectively while oven was operated at 175 ◦ C. All experiments were carried out in triplicate and data were expressed in average values. Experimental data were fitted to the proposed equations and non-linear regression analyses were performed using Newton’s method (Solver, Microsoft Excel 2000, Microsoft Corporation, Redmont, WA, USA). 2.4. Kinetic models Various kinetic models for acid hydrolysis of lignocellulose are available in literature. The model proposed on hydrolysis of cellulose involves polymer glucan of cellulose is degraded to monomer glucose which is subsequently converted to decomposition products [19,20]. The process can be represented as: glucan(s) → glucose → decomposition products The acid hydrolysis of hemicellulose proposed by different authors were, originated from two different kinetic mechanisms [16,17,19–21]. In one mechanism it is assumed that xylan is first converted to xylooligosaccharide, which in turn gives xylose on further hydrolysis and xylose is subsequently decomposed to furfural. Other mechanism does not include the intermediate formation of xylooligosaccharide, where xylan is converted to xylose which in turn is dehydrated to furfural. The accuracy of final results for both the mechanisms is same. The simplest kinetic model for hemicellulose hydrolysis involves pseudohomogeneous irreversible first-order series reactions, which is represented by the following scheme: xylan(s) → xylose(aq) → decomposition products Thus it can be generalized as k1
k2
polymer−→monomer−→decomposition products 2.2. Acid hydrolysis Acid hydrolysis of OPEFB biomass was carried out in 125 ml Erlenmyer flasks. The media consisted of 2–6 g H2 SO4 /100 g liquor using a charge of 1 g OPEFB fiber/8 g liquor on dry basis. Operating temperature of hydrolysis was 120 ◦ C and samples were collected at various time intervals in the range of 0–90 min. The samples were diluted with water and insoluble solids were separated from aqueous solution by filtration. The filtrate was analyzed for xylose, glucose, acetic acid and furfural. 2.3. Analytical methods Xylose and glucose in the acid hydrolysate were analyzed by High Performance Liquid Chromatograph (HPLC, SHIMADZU) using SUPELCOSIL LC-NH2 column and RI detector. Aqueous acetonitrile (75%) was used as mobile phase with flow rate of 1.5 ml min−1 and oven temperature was maintained at 50 ◦ C. Acetic acid and furfural were analyzed by Gas Liquid
(1)
where k1 is the rate of monomer released (min−1 ) and k2 is the rate of monomer decomposed (min−1 ). Based on this reaction model and solving differential equations, monomer concentration (M) as a function of time (t) can be represented by M = [k1 P0 /(k2 − k1 )](e−k1 t − e−k2 t ) + M0 e−k2 t
(2)
where P and M represents concentration of polymer and monomer, respectively. The subscript 0 represents at time t = 0. Assuming M0 to be nearly equal to 0, Eq. (2) can be simplified as M = [k1 P0 /(k2 − k1 )](e−k1 t − e−k2 t )
(3)
The kinetic model for cellulose hydrolysis involves pseudohomogeneous irreversible first order series reactions, which can be represented by the following scheme: k3
k4
glucan(s)−→glucose(aq)−→decomposition products
(4)
S.H.A. Rahman et al. / Biochemical Engineering Journal 30 (2006) 97–103
99
As the operating conditions in the present investigation were not favorable for decomposition of glucose, hence the model can be modified as k3
glucan(s)−→glucose(aq)
(5)
Solving differential equations, concentration of glucose (G) as a function of time can be represented as G = G0 (1 − e−k3 t )
(6)
where k3 is the glucose generation rate (min−1 ) and G0 is the potential glucose concentration which can be estimated by regression analysis. The model for furfural concentration (F) as a function of time can be represented as F = F0 (1 − e−k4 t )
(7)
where F0 is the potential concentration of furfural which can be determined by regression analysis and k4 is the furfural generation rate (min−1 ). Acetyl groups originated from hemicellulose and model for acetic acid generation can be represented by the following scheme: k5
acetyl groups−→acetic acid
(8)
Concentration of acetic acid (A) in the hydrolysate as a function of time can be represented as follows: A = A0 (1 − e−k5 t )
(9)
where A0 is potential concentration of acetic acid which can be obtained by regression and k5 is the acetic acid generation rate (min−1 ). Eqs. (3), (6), (7) and (9) were applied to sulfuric acid hydrolysis of OPEFB fiber. Kinetic parameters and constants were evaluated statistically and determination coefficient R2 of each was obtained in order to find out the reliability of the models. 3. Results and discussion
Fig. 1. Effect of sulfuric acid concentration and reaction time on experimental and predicted concentrations of xylose released for acid hydrolysis of OPEFB fiber at 120 ◦ C.
where Xp0 is the initial amount of xylan polymer present in the OPEFB fiber on dry basis (24.01 g xylan/100 g OPEFB fiber), 150/132 is the stoichiometric factor and LSR is liquid solid ratio (8 g liquid/g of OPEFB fiber). Analysis of acid hydrolysate showed that highest release of xylose from hemicellulose was 31.1 (g/l) when acid concentration was 2% which was 91.2% of potential concentration of xylose. On the other hand with 4 and 6% sulfuric acid, maximum xylose released were 30.7 and 30.27 (g/l), respectively. This is shown in Fig. 1. From the figure it was observed that with increase in acid concentration, xylose concentration attained a peak value for any acid concentration studied and with prolong reaction time, concentration of xylose in the resulting hydrolysate was ultimately decreased. Experimental results suggest that probably there exist some decomposition reaction which leads to dehydration of xylose to furfural. During acid hydrolysis other sugar glucose was also released but in low concentration which is shown in Fig. 2. The highest amount of glucose released in the hydrolysate was 4.1 (g/l) when acid concentration was 6%. On the other hand with 2 and
Analysis of the empty fruit bunch fiber was carried out for determination of principal components using quantitative acid hydrolysis following standard method [18] which is shown in Table 1, expressed on an oven-dry basis. If we assume that polymer xylan is completely converted to xylose without formation of any decomposition products, then P0 can be represented by equivalent amount of xylose by following calculation: P0 = [Xp0 × 150 × 10]/[132 × LSR] = 34.1 g xylose/l (10) Table 1 Main components of oil palm empty fruit bunch fiber (on an oven-dry basis) Main fraction
Composition (%)
Glucan Xylan Lignin (acid insoluble) Ash Others
42.85 24.01 11.70 0.52 20.92
Fig. 2. Dependence of experimental and predicted concentrations of glucose released on acid concentration and reaction time for sulfuric acid hydrolysis of OPEFB fiber at 120 ◦ C.
100
S.H.A. Rahman et al. / Biochemical Engineering Journal 30 (2006) 97–103 Table 2 Kinetics and statistical parameters of components released during H2 SO4 hydrolysis of OPEFB fiber at 120 ◦ C
Fig. 3. Effect of sulfuric acid concentration and reaction time on experimental and predicted concentrations of furfural released for acid hydrolysis of OPEFB fiber at 120 ◦ C.
4% H2 SO4 , maximum glucose released were 1.8 and 3.2 (g/l), respectively. The release of glucose could be either from hemicellulose or cellulose chain. Hence equivalent concentration of glucose from glucan could not be determined. The amount of furfural generated as decomposition product in the resulting hydrolysate is shown in Fig. 3. From the figure it was evident that when H2 SO4 concentration was increased from 2 to 6%, furfural concentration was increased in the resulting hydrolysate. The highest concentration of furfural noticed was 3.3 (g/l) when acid concentration was 6%. This result was in agreement with the reduction of xylose concentration at higher acid concentration which was observed in Fig. 1. It is known that during hydrolysis process acetic acid is generated in the hydrolysate from acetyl groups of hemicellulose. The maximum and minimum concentration of acetic acid in the resulting hydrolysate was found to be 3.93 and 0.59 (g/l) when sulfuric acid concentrations were 6 and 2%, respectively which is shown in Fig. 4.
2% H2 SO4
4% H2 SO4
6% H2 SO4
Xylose k1 (min−1 ) k2 (min−1 ) R2
0.0451 0.0015 0.99
0.0827 0.0029 0.99
0.1695 0.0057 0.99
Glucose k3 (min−1 ) G0 R2
0.0238 3.50 0.99
0.0332 3.70 0.99
0.0518 4.43 0.99
Furfural k4 (min−1 ) F0 R2
0.0031 4.88 0.98
0.0063 6.57 0.99
0.0118 5.28 0.98
Acetic acid k5 (min−1 ) A0 R2
0.0064 6.1 0.96
0.0142 5.55 0.98
0.0189 5.02 0.99
3.1. Kinetic model of xylose concentration Kinetic and statistical parameters obtained on hydrolysis of OPEFB fiber at 120 ◦ C with sulfuric acid in the concentration range of 2–6% is shown in Table 2. Experimental and predicted data for xylose release with various acid concentration and reaction time is shown in Fig. 1. From the table it was evident that with sulfuric acid concentration of 2% xylose generation rate k1 was higher than the decomposition rate k2 which are 0.0451 and 0.0015 (min−1 ), respectively. The determination coefficient R2 showed a good agreement between experimental and predicted data for all regressions. It was also evident that with increase in acid concentration values of k1 and k2 were also increased, which indicated that with increase in acid concentration a shorter reaction time was needed to obtain maximum release of xylose in the resulting hydrolysate. A generalized model for predicting all products was developed by relating kinetic parameters with sulfuric acid concentration by empirical equation kg = k0 Can
Fig. 4. Dependence of experimental and predicted concentrations of acetic acid released on sulfuric acid concentration and reaction time for acid hydrolysis of OPEFB fiber at 120 ◦ C.
(11)
where g is in the range of 1–5; k0 and n the regression parameters; Ca is H2 SO4 concentration expressed in % w/w. Thus generalized model for xylose release is represented by empirical Eq. (12) where k1 is correlated with acid concentration Ca . Similarly xylose decomposition rate k2 is represented by empirical Eq. (13). The determination coefficients R2 for both the parameters were in good agreement which is shown in Table 3. By combining Eqs. (12) and (13) with the models of xylose formation and degradation it is possible to predict xylose concentration at any time and acid concentration within the time period (0–90 min) and acid concentration (2–6%) studied. The generalized model predicted that maximum xylose concentration of more than 29 (g/l) could be obtained with 2, 4 and 6% sulfuric acid at reaction temperature of 120 ◦ C when hydrolysis reaction time were 90, 45 and 15 min, respectively. The dependence of xylose concentration with acid concentration and reaction time is shown
S.H.A. Rahman et al. / Biochemical Engineering Journal 30 (2006) 97–103
101
Table 3 Generalized models for prediction of kinetic parameters of OPEFB fiber hydrolysis with H2 SO4 at 120 ◦ C Products
Models
R2
Xylose
k1 k2 k3 k4 k5
0.97 0.98 0.95 0.98 0.98
Glucose Furfural Acetic acid
= 0.013 Ca1.4217 (12) = 0.0005 Ca1.3849 (13) = 0.0127 Ca0.768 (14) = 0.0011 Ca1.3388 (15) = 0.0037 Ca0.9187 (16)
by response surface in Fig. 5. Thus from the response surface it is possible to select shorter reaction time to obtain maximum release of xylose with minimum decomposition products in the resulting hydrolysate. Again by analyzing different response surfaces it is possible to reach optimum condition of the hydrolysis process. 3.2. Kinetic model of glucose concentration In the model of glucose release (Eq. (6)), G0 could not be obtained experimentally as glucose could have released from both hemicellulose and cellulose. Hence G0 and kinetic parameter k3 were obtained by regression and determination coefficient R2 showed was in good agreement with experimental and predicted data. The values of G0 obtained were in the range of 3.5–4.43 (g/l). The values of k3 were found to be in the range of 0.0238–0.0518 (min−1 ), which is shown in Table 2. It was observed that the values of k3 were increased with increase in acid concentration. Hence a generalized model for prediction of glucose concentration was developed which was correlated with acid concentration and kinetic parameter k3 . This is presented by empirical Eq. (14) and determinant coefficient R2 was well fitted which is given in Table 3. From the table it was observed that the value of regression parameter n for glucose generation
Fig. 6. Effect of sulfuric acid concentration and reaction time on generalized model for prediction of glucose concentration.
rate k3 was 0.768 which was lower than that of corresponding value of regression parameter (n = 1.4217) for xylose generation k1 . It is known that hemicellulose is amorphous in character whereas cellulose is crystalline in nature. Moreover hydrolysis of cellulose is strongly dependent on degree of crystallinity and swelling state of cellulose [22]. The results suggested that probably heterogeneous reaction occurred during release of glucose from glucan. Hence the effect of acid on glucose generation rate k3 was not similar in nature compared to that of effect of acid on xylose generation rate k1 . Experimental and predicted data of glucose generation in the resulting hydrolysate is shown in Fig. 2. Thus with the help of Eq. (14) and glucose generation model, it is possible to predict glucose concentration within the experimental range studied. The models predicted highest glucose concentration of 4.38 (g/l) at most severe reaction conditions (6% acid concentration and 90 min reaction time) and greater than 1.0 (g/l) could be obtained if the reaction was carried out for time period (t < 30 min) for any acid concentration studied. The response surface of the generalized model for glucose generation with increase in acid concentration and reaction time is shown in Fig. 6. From the response surface it was observed that there was no decomposition reaction occurred during the hydrolysis process. To obtain maximum xylose concentration in the resulting hydrolysate, it is necessary to keep glucose concentration as low as possible. Hence response surface in this respect can help to conduct the experiment in the range (shorter reaction time) where glucose concentration can be kept at low level for maximizing xylose concentration in the resulting hydrolysate. 3.3. Kinetic model of furfural concentration
Fig. 5. Effect of sulfuric acid concentration and reaction time on generalized model for prediction of xylose concentration.
In the acid hydrolysis of OPEFB biomass furfural is the principal degradation product of xylose. Kinetic and statistical
102
S.H.A. Rahman et al. / Biochemical Engineering Journal 30 (2006) 97–103
Fig. 7. Effect of sulfuric acid concentration and reaction time on generalized model for prediction of furfural concentration.
parameters obtained for furfural is shown in Table 2. It was found that values of F0 and kinetic parameter k4 were within the range of 4.88–6.57 (g/l) and 0.0034 to 0.0118 (min−1 ), respectively. The determinant coefficients R2 were well fitted with furfural generation model. Experimental and predicted data is shown in Fig. 3. It was noted that furfural formation rate k4 was higher than xylose degradation rate. It was also noted that kinetic parameter k4 was increased with increase in acid concentration. A generalized model was developed to correlate kinetic parameter k4 with acid concentration for prediction of furfural concentration at any time and acid concentration within the operating range studied. The empirical Eq. (15) represents the generalized model which is shown in Table 3. The graph of generalized model for furfural formation is shown in Fig. 7. From the response surface it was observed that with increase in reaction time and acid concentration, furfural concentration in the resulting hydrolysate was increased. It is known that furfural is an inhibitory compound to fermentation process and hence its concentration in the hydrolysate should be minimized to facilitate optimum use of hydrolysate for xylitol production. From the response surface it can be interpreted that it is better to conduct hydrolysis process with higher acid concentration and lower reaction time to minimize the formation of furfural in the resulting hydrolysate. In this case generalized model can help to predict concentration of furfural and hence selection of acid concentration and reaction time can be done in order to obtain maximum concentration of xylose in the hydrolysate while keeping furfural concentration at minimum level. 3.4. Kinetic model of acetic acid concentration In acid hydrolysis of OPEFB fiber, acetic acid is mainly released from acetyl groups of hemicellulose. Kinetic and statistical parameters obtained for acetic acid generation is shown in Table 2. Experimental and predicted data were well fitted
Fig. 8. Effect of sulfuric acid concentration and reaction time on generalized model for prediction of acetic acid concentration.
with the model as is evident from R2 value. Experimental and predicted data for the acid hydrolysis of OPEFB fiber at reaction temperature of 120 ◦ C is shown in Fig. 4. It was observed that value of A0 and k5 were within the range of 6.1–5.02 and 0.0064–0.0189 (min−1 ), respectively. It was also observed that the value of kinetic parameter k5 was increased with increase in acid concentration. A generalized model was developed to correlate k5 with acid concentration for prediction of acetic acid concentration within the experimental range of reaction time (0–90 min) and acid concentration (2–6%) studied. This is presented by empirical Eq. (16) as shown in Table 3. The value of regression parameter n for acetic acid generation k5 obtained was 0.9187 which was lower than the value of regression parameter (n = 1.4217) for xylose generation rate k1 . It is reported that effect of acid on acetyl removal from hemicellulose is lower compared to that of effect of acid on xylan removal which may be the reason for obtaining different n values [23]. The response surface of generalized model for acetic acid released in the hydrolysate is shown in Fig. 8. From the figure it can be interpreted that acetic acid concentration was increased with increase in sulfuric acid concentration and reaction time. Hence to maximize xylose concentration in the resulting hydrolysate it is better to conduct the experiment at high sulfuric acid concentration and low reaction time so that concentration of acetic acid in the hydrolysate can be kept at low level. The generalized models of xylose, glucose, furfural and acetic acid on dilute sulfuric acid hydrolysis of OPEFB fiber predicted several options for obtaining high xylose concentration in the resulting hydrolysate. Thus when the hydrolysis was conducted with 2% sulfuric acid at reaction temperature of 120 ◦ C for 90 min reaction, then xylose, glucose, furfural and acetic acid predicted in the resulting hydrolysate were 29.96, 2.99, 1.08 and 2.82 (g/l), respectively. Whereas when the experiment was conducted with 4% acid for 45 min reaction time the concentration of xylose, glucose, furfural and acetic acid predicted
S.H.A. Rahman et al. / Biochemical Engineering Journal 30 (2006) 97–103
in the resulting hydrolysate were 29.83, 2.99, 1.78 and 2.49 (g/l), respectively. On the other hand when the reaction was carried out with 6% acid for 15 min of reaction time the corresponding concentrations of xylose, glucose, furfural and acetic acid in the acid hydrolysate predicted were 29.4, 2.34, 0.87 and 1.25 (g/l), respectively. Hence the optimum operating condition for production of xylose from OPEFB fiber is to conduct the hydrolysis process at 120 ◦ C with 6% sulfuric acid for 15 min reaction time where all the byproducts can be controlled to minimum level. 4. Conclusion Oil palm empty fruit bunch fiber is a palm oil mill waste. It contains 24% xylan, which is a promising source of xylose. Acid hydrolysis of OPEFB fiber was conducted at operating temperature 120 ◦ C under various H2 SO4 acid concentration (2–6%) and reaction time (0–90 min). The maximum concentration of xylose, glucose, furfural and acetic acid obtained in the resulting hydrolysate were 31.1, 4.1, 3.3 and 3.93 (g/l), respectively. Kinetic parameters of mathematical models were determined in order to predict concentration of xylose, glucose, furfural and acetic acid in the resulting hydrolysate and to optimize the hydrolysis process. The optimum acid concentration and reaction time obtained under operating temperature 120 ◦ C were 6% and 15 min, respectively. The optimum concentration of xylose, glucose, furfural and acetic acid released in the hydrolysate were 29.4, 2.34, 0.87 and 1.25 (g/l), respectively. The product Xylose can further be converted to xylitol by bioconversion process. Thus with optimum use of OPEFB fiber, it will not only solve the environmental pollution problem but also will give back a high value product to the palmoil industry. Acknowledgement The authors are grateful to the Ministry of Science, Technology and Environment, Malaysia Government for the financial support of this work under long term IRPA grant (Project: 0302-05-2215 EA008). References [1] R.C. Kuhad, A. Singh, Current and future prospects, Crit. Rev. Biotech. 13 (1993) 151–172. [2] C.E. Wyman, Ethanol from lignocellulosic biomass: Technology, economics and opportunities, Bioresour. Technol. 50 (1994) 3–16. [3] J.B.A. Silva, I.M. deMancilha, M.C.D. Vannetti, M.A. Teixeira, Microbial protein production by Paecilomyces variotii cultivated in eucalyptus hemicellulosic hydrolysate, Bioresour. Technol. 52 (1995) 197–200.
103
[4] J.C. Parajo, H. Dominguez, J.M. Dominguez, Production of xylitol from raw wood hydrolysate by Debaryomyces hansenii NRRL Y-7426, Bioprocess Eng. 13 (1995) 125–131. [5] T. Pepper, P.M. Olinger, Xylitol in sugar-free confections, Food Technol. 10 (1988) 98–106. [6] K.K. Makinen, Dietary prevention of dental caries by xylitol-clinical effectiveness and safety, J. Appl. Nutr. 44 (1992) 16–28. [7] L. Hyvonen, P. Koivstoinen, Food technological evaluation of xylitol, Adv. Food Res. 28 (1982) 373–403. [8] I.C. Roberto, M.G.A. Felipe, I.M. Mancilha, M. Vitolo, S. Sato, S.S. Silva, Xylitol production by Candida guilliermondii as an approach for the utilization of agro industrial residues, Bioresour. Technol. 51 (1995) 255–257. [9] J.M. Dom´ınguez, N. Cao, C.S. Gong, G.T. Tsao, Dilute acid hemicellulose hydrolysates from corn cobs for xylitol production by yeast, Bioresour. Technol. 61 (1997) 85–90. [10] G. Garrote, H. Dominguez, J.C. Parajo, Generation of xylose solutions from eucalyptus wood by autohydrolysis-posthydrolysis processes: posthydrolysis kinetics, Bioresour. Technol. (2001) 155–164. [11] B.P. Lavarack, G.J. Griffin, D. Rodman, The acid hydrolysis of sugarcane bagasse hemicellulose to produce xylose, arabinose, glucose and other products, Biomass Bioenerg. 23 (2002) 367–380. [12] S.S. Silva, M.G.A. Felipe, J.B.A. Silva, A.M.R. Prata, Acid hydrolysis of eucalyptus grandis chips for microbial production of xylitol, Process Biochem. 33 (1998) 63–67. [13] M. Neureiter, H. Danner, C. Thomasser, B. Saidi, R. Braun, Dilute acid hydrolysis of sugar cane bagasse at varying conditions, Appl. Biochem. Biotechnol. 98 (2002) 49–58. [14] C. Van Zyl, B.A. Prior, J.C. Du Preez, Acetic acid inhibition of dxylose fermentation by Pichia stripitis, Enzyme Microb. Technol. 13 (1991) 82–86. [15] J.C. Parajo, V. Santos, F. del Rio, Hydrolysis of the cellulosic fraction of pinus pinaster wood. II. Production at higher pressure than atmospheric one, Afinidad 52 (1995) 267–274. [16] G. Marton, J. Dencs, L. Szokonya, Principles of Bio Mass Refining: in Handbook of Heat and Mass Transfer, Gulf Publishing, Houston, 1989, 609–652. [17] S.J. Tellez-Luis, J.A. Ramirez, M. Vazquez, Mathematical modelling of hemicellulose sugar production from sorghum straw, Jl. Food Engg. 52 (2002) 285–291. [18] Chemical Analysis and Testing Standard Procedure, National Renewable Energy Laboratories, Golden, Co., NREL (1996), No. 002-004. [19] J.F. Saeman, Kinetics of wood saccharification, hydrolysis of cellulose and decomposition of sugars in dilute acid at high temperature, Ind. Eng. Chem. 37 (1945) 43–52. [20] D.F. Root, J.F. Saeman, J.F. Harris, Chemical conversion of wood residues. Part II. Kinetics of acid-catalyzed conversion of xylose to furfural, Forest Prod. J. 9 (1959) 158–165. [21] F. Carrasco, C. Roy, Kinetic study of dilute-acid-pre-hydrolysis of xylan containing biomass, Wood Sci. Technol. 26 (1992) 189–208. [22] Q. Xiang, Y.Y. Lee, P.O. Pettersson, R.W. Torget, Heterogeneous aspects of acid hydrolysis of ␣-cellulose, Appl. Biochem. Biotechnol. 105–108 (2003) 505–514. [23] J.F. Harris, A.J. Baker, A.H. Connor, T.W. Jeffries, J.L. Minor, R.C. Petersen, R.W. Scott, E.L. Springer, T.H. Wegner, J.I. Zerbe, General technical report FPL-45, US Department of Agriculture Forest Products Laboratory, Madison, WI, 1985.