BIOETHANOL PRODUCTION FROM SAGO STARCH Maizirwan Mel, Husna Bt Muhammad Nadzri, Mohd Hider Kamarudin, Mohd Ismail Abd Karim Bioprocess and Molecular Engineering Research Unit, Faculty of Engineering, International Islamic University Malaysia, P.O. Box 10, 50728, Kuala Lumpur, Malaysia Corresponding Email: maizirwan@iium.edu.my Abstract Two stage processes prior to production has been proposed which is hydrolysis and fermentation. Commercial enzymes were used in this two steps hydrolysis; α-amylase (liquefaction step) and glucoamylase (saccharification step). Optimization was carried out in both stages; hydrolysis and fermentation. Three parameters are involved in optimization of % (DE); sago starch concentration (20% (w/v), 0% (w/v), 40% (w/v)), glucoamylase enzyme (52 U/g, 78 U/g, 104 U/g) and time during saccharification (1, 2, hours). Three parameters are involved in optimization of ; agitation (100 rpm, 150 rpm, 200 rpm), inoculums (1% (v/v), % (v/v), 5% (v/v)) from constant initial stock of 2.5x10 6 cells/ml and ph (4, 5, 6). Both optimization studies were carried out using Box-Behnken design. The experiment showed that the optimum parameters for hydrolysis study was identified to be glucoamylase (75.87 U/g), substrate concentration (28.49% (w/v)), and time (2 hours) which produced 62.15 g/l glucose as the fermentation substrate. For fermentation study, it was identified that the optimum parameters that produced 29.25 g/l were 167 rpm agitation,.4% (v/v) inoculums and ph 5. Sago; fermentation; hydrolysis; ; optimization I. INTRODUCTION Due to the scarce resources and increasing environmental concern, replacing the non renewable energy with alternative resources is one of the most studied opportunities nowadays. One way is by utilizing the solar energy in form of biomass to produce also called as bio [1]. Balat et. al [2] pointed out that an appropriate blending of with conventional fuels has proven not only in promoting a cleaner environment but also helps balancing the economical value of fuel price. Bio is mainly produced from sugar base crops, starchy and cellulosic raw materials via fermentation process. Ethanol from sugar base crops such as sugar cane can be converted directly since the sugar is already in fermentable form. For starchy raw materials, hydrolysis must be conducted first before proceeding with fermentation process []. Hydrolysis process can be divided into two parts; liquefaction and saccharification. Liquefaction process utilizes α-amylase enzyme, whereas saccharification process utilizes glucoamylase enzyme [].This is an early process to produce sugar in a readily fermentable form. One type of starchy crops that has promising future in bio is sago palm or scientifically known as Metroxylon sagu [4]. As cited in [4], sago palm is also known as the starch crops of the 21 st century by most scientists. This proves that sago palm has brighter prospect as the source for carbohydrate as well as for bio production. Starch from sago palm is the only commercial starch source that is derived from stem and contains bulky amount of starch in its trunk [5]. Karim et al [5] also stated that, sago palm plantation mainly concentrated in most tropical countries. In Malaysia, approximately 12% of Sarawak total area is mainly covered with sago plantation. Generally, good production of bio depends on the availability to produce large concentrations and maintaining high quality yield at the end of the process [6]. Ratnam et al [6] pointed out that, substrate selection is the main cost factor for industry. Traditionally, fermentation greatly depends on sugar rich substrates such as sugar cane since they are already in form of readily fermented carbohydrate [7]. Nevertheless, sugar cane is costly and difficult to obtain since it is categorized under seasonal crops [7]. Therefore, it is a great economic advantage to expand the substrate choice to either starchy crops and cellulosic materials which are cheaper compared to sugar rich substrate [7]. Seeing it as the potential for carbohydrate sources, the study on sago as the source for industry should be done intensively. Starchy crops primarily involved for production includes cereal grains (corn, wheat, barley, sorghum), tuber roots (potato), roots (cassava) and legumes (peas, beans) [8]. Due to the availability and lack of involvement of sago for production, it is a challenging opportunity to study and optimize the parameters involved in producing high level of product with excellent yield. This project involved two stages of bioconversion namely, hydrolysis and fermentation using sago starch as the raw material. In general, hydrolysis process is carried out to convert the long polymer like chain of carbohydrate into monomers of glucose. In this study, two steps of enzymatic hydrolysis process which involves liquefaction and saccharification was carried out. High concentration of depends on the availability substrate []. Therefore, optimization of glucose yield is necessary to improve the overall production. Here, the optimization process of glucose yield focuses at saccharification step of hydrolysis. There are three parameters involved which include substrate, glucoamylase enzyme and time. After that, fermentation process is carried out by optimizing the parameter involved in the process. Three parameters were analyzed at fermentation stage which includes ph, agitation and inoculums size. All experiments are conducted in shake flask in Bioprocess Engineering Laboratory of IIUM.
The initial studies from laboratory level can be further used in upscale level of bio industry. Mojovic et al [9] stated that US and Brazil are the major contributors toward industry. Perhaps, studying sago as an fermentable substrate will be able to break the monopoly of and provides better opportunity for Malaysia in improving their economy. In addition, the abundance amount of sago palm provides cheap selection of subtract that will subsequently reduce the overall cost production of. Moreover, this will also solve the over supply of agricultural resources and offers a great opportunity in helping farmers in contributing towards agricultural industry. Hence, diverging part of sago crop to production does not only help in reducing the fuel price but also maintaining the price of sago derived products [10]. Furthermore, it will also provide agricultural sustainability for sago palm since it is fully utilized not only for food, but also energy consumption [11]. The aims of this study were to optimize the process condition for hydrolysis of sago starch to glucose with α- amylase and glucoamylase enzyme and to optimize fermentation process of saccharified sago starch to using Saccharomyces cerevisiae in shake flask experiment. A. Substrate II. MATERIALS AND METHODS Sago starch (powdered form containing 9% dry weight starch content) was obtained from Sarawak Craun Research Institute. Initially, 100 ml beaker was filled with 20, 0 and 40 g sago starch (powdered form) and added with 100 ml distilled water (to make a total concentration of 20% (w/v), 0% (w/v) and 40% (w/v)).the resulting slurry was heated to 80ºC for 15 minutes for starch gelatinization [7]. B. Enzymes α-amylase for liquefaction was used and produced from Bacillus subtilis with an activity of 25,000 U/mL. One unit of α-amylase equals to the amount of enzyme which liquefies soluble starch to get 1 mg dextrins at 70ºC and ph 6.0 in one minute. Glucoamylase for saccharification was also used and produced from Aspergillus niger with an activity of 10,000 U/mL. One unit of glucoamylase equals to the amount of enzyme which hydrolyzes soluble starch to 1 mg glucose at 40ºC and ph4.5 in one hour. C. Hydrolysis Initially, 100mL beaker was filled with 20, 0 and 40 g sago starch (powdered form) according and added with 100 ml distil water (to make a total concentration of 20% (w/v), 0% (w/v) and 40% (w/v)).the resulting slurry was heated to 80ºC for 15 minutes for starch gelatinization [7]. After gelatinization, 25 U/g of α-amylase was added to the slurry and heated at 80ºC using hot plate with mixing for 1 hour using magnetic stirrer [12]. Then, the temperature was lowered down to 50ºC for minutes to make sure that the overall temperature has already drop in preparation for saccharification to take place. Then, the desired concentration of glucoamylase (52 U/g, 78 U/g, and 104 U/g) was added and left to react within the specified hours (1, 2, hours). Mixing was carried out throughout the whole reaction using a magnetic stirrer. D. Analyses After hydrolysis, all samples were collected and cooled down to 5ºC. Then, centrifugations of samples were carried out at 5000 rpm for 0 minutes. The supernatant was collected and reducing sugar was analyzed using dinitrosalicylic acid (DNS) method [1]. Then % (DE) is calculated based on equation (1) below according to [14]. % DE=g reducing sugar expressed as glucose x 100% (1) g dry solid weight sago starch % DE is calculated to determine the effective conversion process to glucose based on the amount of sago starch. After fermentation, the samples of 10 ml were collected every 8 hours of fermentation and centrifuged at 5000 rpm for 15 minutes at 4ºC to remove the cell debris. The supernatants left were then used for and reducing sugar analysis. The reducing sugar analysis was carried out based on [1] during hydrolysis. For determination, two milliliters of the supernatant was pipette to a test tube, followed by the addition of 2- drops of potassium dichromate solution (dissolve 2 g of potassium dichromate in 80 cm³ of distilled water and carefully add 10 cm³ of concentrated sulphuric acid). The tube was vigorously swirled in order to ensure homogenous suspension. The tube then was placed in a water bath at 60ºC for about 20 minutes. Color development was observed and the color was left to stable for at least 0 minutes at room temperature. The color density was measured at 600 µm wavelength by using the spectrophotometer with the blank being distilled water. The content of the samples will be estimated by reference to the standard curve E. Inoculums Preparations One colony of S. cerevisiae was isolated and inoculated into 100 ml shake flask containing 50 ml of yeast malt under sterilized condition. Then, it was incubated for 24 hours at 0ºC and 150 rpm. The inoculums size was prepared to have an initial concentration of 2.5x10 6 cells per ml. Dilutions are made if the concentration is too high. F. Fermentation For medium preparation, 0.05% (w/v) of urea and 0.05% (w/v) of NPK (nitrogen, phosphorus and potassium) were added to 250 ml shake flask containing 100 ml of saccharified sago starch [12]. The glucose concentration for the fermentation media was 62.15 g/l which was obtained from the hydrolysis process using the optimum condition. The ph was adjusted to ph 4, 5 and 6 by the addition of required volume of 1M NaOH and 1M H 2SO 4. Then, the inoculated yeast (1 %v/v, %v/v, 5 %v/v) was added to the medium from the same inoculums stock (2.5x10 6 cells per ml). Then, the sample was transferred into the shaker and agitation speed was set to 100, 150 and 200 rpm with fixed temperature of 0º C for 72 hours [12]. Fermentation process was carried out under sterilized condition. Readings were taken every 12 hours of fermentation for 72 hours and recorded.
G. Statistical Analysis Analysis of the results was done statistically with the aid of Design Expert software v 8.01. The analysis was done based on Ratnam s method [6]: Analysis of Variance (ANOVA) is used to investigate and model the relationship between a response variable with one or more independent variable. Response Surface Methodology (RSM) uses quantitative data from appropriate experiments to simultaneously determine and solve multi variant equations. The factors that should be taken into consideration: i) system is well understood where all the critical factors are known ii) region of interest where the factor levels influencing product is known iii) factors vary continuously throughout experimental range has been tested iv) a mathematical function relates the factors to the measured response and v) the response defined by the function is a smooth curve. From the designed experiment, 15 runs have been done for the optimization of the parameters during hydrolysis and fermentation. The model parameters during both optimization processes will be: Y= f (A,B,C) (2) Y=b 0+b 1A+b 2B+b C+b 11A 2 +b 22B 2 +b C 2 +b 12AB +b 2BC+b 1AC () Where Y is the output (glucose (% glucose conversion), ), predicted response; A, B and C the independent variables, b 0 the offset term, b 1, b 2, b the linear effects, b 11, b 22, b the squared variables and b 12, b 2, b 1 are the interaction effects. III. RESULTS AND DISCUSSION A. Design of experiment and Statistical Analysis for Hydrolysis For the optimization process of hydrolysis, three factors namely; sago starch concentration (A), glucoamylase (B) and time (C) were observed to study the most effective conversion of sago starch to glucose. The overall optimization study result of hydrolysis is shown in Table 2. In order to search for the optimum parameter for % yield, a total of 15 treatments for each optimization were established using Design Expert Software with applying Box-Behnken Design with different range of the said parameters. The quadratic relating to the % (DE) with independent variables A, B and C are as follows: Y (%) = -69.409 + 1.177A + 1.0216B + 2.95647C - 0.01017A 2-0.029489B 2-5.05582C 2 + 0.00865894AB + 0.000580769AC - 0.14128BC (4) The average triplicate value of % (DE) was referred as dependent variable or response of Y. Table 1 shows the predicted levels of response Y using equation (4) were given along with actual values. TABLE 1. Actual and predicted value of % (DE) Std. Order (%DE) predicted value 2 20.00 1 10.00 Glucoamylase (A) U/g starch Factors Substrate (B) (%w/v) Time (C) (h) predicted vs actual R 2 = 0.9799 0.00 0.00 10.00 1 20.00 2 actual value Actual value %DE Predicted value %DE 1 52 20 2 17.40 16.55 2 104 20 2 10.65 10.4 52 40 2 10. 10.56 4 104 40 2 12.59 1.44 5 52 0 1 11.09 11.92 6 104 0 1 10.07 10.27 7 52 0 11.18 10.97 8 104 0 10.22 9.8 9 78 20 1 14.4 14.45 10 78 40 1 16.85 15.79 11 78 20 15.0 16.6 12 78 40 12.07 12.05 1 78 0 2 21.90 22.67 14 78 0 2 2.40 22.67 15 78 0 2 22.70 22.67 (%DE) Figure 1. Parity plot showing the distribution of predicted versus actual values of % DE Figure 1 proves that a satisfactory correlation coexists between the actual and predicted value, wherein, the points cluster around the linear line which indicates a good fit model. From the ANOVA analysis, the determination coefficient R 2 was evaluated as to test fit of the design experiment. The mathematical adjust of those values generated a R 2 = 0.9799, revealing that the model could not explain only 2.01% of the overall effects, showing that it is a robust statistical model [15]. Table 2 presented the corresponding analysis of variance (ANOVA).
(%) (%) TABLE 2. ANOVA for Response Surface Quadratic Model during hydrolysis Source Sum of Squares DF Mean Square F Value Prob > F Model 11.8 9 4.60 27.0 0.0010 significant A 5.25 1 5.25 4.10 0.0988 X = A: glucoamylase Y = B: substrate C: time = B: substrate (%w/v) 40.00 0.00 (%) 20.6994 18.6448 18.6448 B 4.42 1 4.42.45 0.1222 2 C 1.69 1 1.69 1.2 0.022 A 2 179.60 1 179.60 140.1 0.0001 significant B 2 2.11 1 2.11 25.08 0.0041 significant C 2 94.8 1 94.8 7.74 0.0004 significant AB 20.27 1 20.27 15.84 0.0105 significant AC 0.000912 1 0.000912 0.0007125 0.9797 BC 7.98 1 7.98 6.24 0.0547 Residual 6.40 5 1.28 Lack of Fit 5.28 1.76.1 0.251 Pure Error 1.12 2 0.56 Cor Total 17.78 14 not significant To determine the significance of the overall model and individual model term, it depends on the value of Prob. > F. Prob. > F less than 0.0500 indicates that the model is significant. From the result, the overall ANOVA for response surface for quadratic model demonstrates a significant model with value of Prob > F= 0.0010. Based on Table 4.2, A 2 (0.0001), B 2 (0.0041), C 2 (0.0004) and AB (0.0105) are significant model terms. Other than Prob. > F, F-test also can be used to determine the significance of the individual model terms. As the F value >10, indicates the model terms are significant. From Table 4.2, A 2, B 2, C 2 and AB show significant model terms with F value 140.1, 25.08, 7.74 and 15.84 respectively. The Lack of Fit-F value implies the Lack of Fit is not significant relative to the pure error. There is a 25.1% chances that the Lack of Fit-F value this large could occur due to noise. Based on the ANOVA analysis, it can be concluded that the significant variables with largest effect on % are squared term of glucoamylase enzyme (A 2 ), squared term of sago starch concentration (B 2 ) and squared term of time (C 2 ) and interaction between glucoamylase and sago starch (AB). Other variables which excluded from the said variables do not have great interaction towards % (DE). To investigate the effects of each independent variable, the D response surface and contour plots response surface are used. These two techniques are the graphical representation of regression equation in order to determine the optimum values of variables [16]. The maximum predicted value is referred by the surface defined in the smallest ellipse in the contour diagram. The perfect interaction between the independent variables can be shown when elliptical contours are obtained [17]. The graph representations are shown in Figure 2 to Figure 4. X = A: glucoamylase Y = B: substrate C: time = 22.7541 19.6721 1.5082 (%w/v) Figure 2. Response surface described by the model equation to estimate % value over independent variables substrate concentration % (w/v) and glucoamylase enzyme U/g. (i) Upper: Contour plot surface response (ii) Lower: D plot surface response. 10.4262 X = A: glucoamylase Y = C: time B: substrate = 0.00 X = A: glucoamylase Y = C: time 20.00 40.00 (hours) C: time 22.678 B: substrate = 0.00 5 6 78.00 9 10 19.54 16.00 12.706 9.8228 2.50 1.50 0.00 B: substrate (hours) A: glucoamy lase (U/g) (U/g) (%) 20.6994 18.6448 12.4808 5 6 78.00 9 10 2.50 C: time 2 20.00 A: glucoamy lase (U/g) Figure. Response surface described by the model equation to estimate % value over independent variables time (hrs) and glucoamylase enzyme (U/g). (i)upper:contour plot surface response(ii) Lower: D plot surface response 1.50 5 6 A: glucoamylase 5 78.00 9 78.00 10 9 6 A: glucoamylase (U/g) 10
(%) X = B: substrate Y = C: time A: glucoamylase = 78.00 C: time (hours) 2.50 (%) 18.6448 TABLE. Actual and predicted value of after 72 hours Std. Order Agitation (A) RPM Factors Inoculums (B) ph (C) Actual value g/l Predicted value g/l 1 100 1 5 9.7 10.58 1.50 20.6994 18.6448 20.00 2 0.00 40.00 2 200 1 5 16.02 16.2 100 5 5 14.6 14.15 4 200 5 5 21.54 20.66 X = B: substrate Y = C: time 22.7119 A: glucoamylase = 78.00 20.0455 17.79 14.7126 12.0461 B: substrate (%w/v) 5 100 4 20.04 19.06 6 200 4 17.67 17.7 7 100 6 9.07 9.7 8 200 6 22.25 2.2 9 150 1 4 16.7 16.8 10 150 5 4 18. 19.48 C: time (hours) B: substrate (%w/v) Based on the three figures of surface responds and contour plots, it is observed that maximum was found to be between 2 to 2.5 hours, glucoamylase enzyme between 65 U/g to 78 U/g and substrate concentration 25% to 0% (w/v) in order to obtain an optimum of 22.76% dextrose equivalent value for this experiment. From the optimization study of hydrolysis, it is suggested by Design Expert software that the combination process between glucoamylase (75.87 U/g), substrate concentration (28.49% w/v), and time (2 hours) will produce an optimum degree of hydrolysis of 22.76% dextrose equivalent. B. Design of experiment and statistical analysis for fermentation 2.50 Figure 4. Response surface described by the model equation to estimate % value over independent variables time (hrs) and substrate (%w/v). (i) Upper: Contour plot surface response (ii) Lower: D plot surface response In optimization of fermentation study there are three parameters involved; agitation (A), inoculums (B) and ph (C) in determination of optimum yield. In order to search for the optimum combination parameters, a total of 15 treatments for each optimization were established using Design Expert Software with applying Box-Behnken Design at different range of the said parameters. The quadratic relating to the production of with independent variables A, B and C are as follows: Y ( g/l) = -51.2887+0.255675 A+6.4615 B +17.62125C-0.00197A 2-1.2448B 2.125C 2 +0.00215AB+0.07775AC-0.625BC (5) The average triplicate value of production after 72 hours sampling was referred as dependent variable or response of Y. Table shows the predicted levels of response Y using equation (5) were given along with actual values. 1.50 20.00 2 0.00 40.00 predicted value 11 150 1 6 14.75 1.57 12 150 5 6 19.01 18.91 1 150 5 28.09 25.0 14 150 5 22.59 25.0 15 150 5 25.22 25.0 0 25 20 15 10 5 0 predicted vs actual value R 2 = 0.944 0 5 10 15 20 25 0 actual value Figure 5. Parity plot showing the distribution of predicted versus actual values of yield after 72 hours The parity plot which is in figure 5, shows a satisfactory correlation coexists between the actual and predicted values, wherein, the point s lines within the linear line which indicates a good fit model. From the ANOVA analysis, the determination coefficient R 2 was evaluated as to test fit of the design experiment. For this study, the model shows a good determination coefficient R 2 =0.944 which implies 94.4% of the sample variation in the production is attributed to the independent variables; agitation, inoculums and ph. The R 2 value also indicates that about 5.6% of the variation is not explained by the model [15]. Table 4 shows the corresponding analysis of variance (ANOVA) is presented in.
TABLE 4. ANOVA for Response Surface Quadratic Model during fermentation Source Sum of Squares DF Mean Square F Value Prob > F Model 65.8 9 40.60 9.62 0.0120 significant A 7.87 1 7.87 17.06 0.0091 significant B 2.04 1 2.04 7.89 0.0419 significant C 7. 1 7. 1.691 0.2501 X = A: agitation Y = B: inoculum C: ph = B: inoculum 2.4159 A 2 89.29 1 89.29 20.591 0.0062 significant 15.7144 B 2 91.48 1 91.48 21.096 0.0059 significant 1.1472 C 2 6.06 1 6.06 8.15 0.044 significant AB 0.18 1 0.18 0.04 0.8445 AC 60.45 1 60.45 1.941 0.015 significant BC 1.81 1 1.81 0.417 0.5468 Residual 21.68 5 4.4 Lack of Fit 6.55 2.18 0.288 0.841 Pure Error 15.1 2 7.57 Cor Total 87.06 14 Not significant The probability P-value was also relatively low (P model> F =0.012) indicates the significance of the model where significance is judge on Prob > F less than 0.0500. Meanwhile, model terms with values of Prob > F greater than 0.1000 indicate the model terms are not significant. Based on Table 4.4, A (0.0091), B (0.0419), A 2 (0.0062), B 2 (0.0059), C 2 (0.044) and AC (0.015) are significant model terms. The Lack of Fit-F value implies the Lack of Fit is not significant relative to the pure error. There is 8.41% chances that the Lack of Fit-F value this large could occur due to noise. Based on the ANOVA analysis, it can be concluded that the significant variables with largest effect on % dextrose equivalent are agitation (A), inoculums (B), squared term of agitation (A 2 ), squared term of inoculums (B 2 ), squared term of ph (C 2 ) and interaction of agitation and ph (AC).Other variables which excluded from the said variables do not have great interaction production. The D response surface and contour plots response surface are used to further investigate the effects of each independent variable. The maximum predicted value is referred by the surface defined in the smallest ellipse in the contour diagram. The graph representations are shown in Figure 6 to Figure 8. Based on the three plots, an optimum production which was about 26 g/l was found to achieved at an approximate parameter of ph 5, 167 rpm and % (v/v) inoculums from 62.15 g/l of saccarified sago starch by two stages bioconversion processes (hydrolysis and fermentation). Where, approximately 41.8% of substrate was converted into. The yield was slightly lower compared to the theoretical maximum yield of stated by Chiaramonti [] where from 1 kg of mash can yield approximately 0.5111 kg of. In Ratnam s study [6], an optimum production of 70.68 g/l was produced from 140 g/l sago starch with agitation speed 200 rpm, ph5 and 5% (v/v) inoculums which is about 50.48% of the sago starch was converted into using simultaneous saccharification and fermentation with glucoamylase (AMG) and Zymomonas mobilis. X = A: agitation Y = B: inoculum C: ph = Figure 6. Response surface described by the model equation to yield value over independent variables agitation (rpm) and inoculums % (v/v). (i) Upper: Contour plot surface response (ii) Lower: D plot surface response. X = A: agitation Y = C: ph B: inoculum = X = A: agitation Y = C: ph C: ph 25.7691 B: inoculum = 21.6702 17.5714 25.981 22.12 1.4726 14.408 9.775 10.58 6.00 6.00 5.50 4.50 100.00 12 150.00 17 200.00 1.1472 15.7144 2.4159 (rpm) A: agitation (rpm) 12 A: agitation 100.00 (rpm) Figure 7. Response surface described by the model equation to yield value over independent variables agitation (rpm) and ph. (i) Upper: Contour plot surface response (ii) Lower: D plot surface response. (rpm) 100.00 12 150.00 17 200.00 5.50 C: ph B: inoculum 4.50 A: agitation 12 A: agitation 100.00 150.00 150.00 17 17 200.00 200.00
X = B: inoculum Y = C: ph A: agitation = 150.00 X = B: inoculum Y = C: ph Figure 8. Response surface described by the model equation to yield value over independent variables ph and inoculums % (v/v). (i) Upper: Contour plot surface response (ii) Lower: D plot surface response. C. Validation The fermention parameters were validated to prove the optimum parameters suggested by Design Expert software. Table 5 shows the treatment process was done using three different conditions and the observed yield values were compared to predicted values for confirmation of the predicted values. TABLE 5. Validation results using optimize parameter. Agitation (rpm) C: ph 25.5522 A: agitation = 150.00 22.5557 Factors Inoculums 6.00 5.50 4.50 15.7144 ph 2.4159 B: inoculum B: inoculum Ethanol Concentration g/l Actual Predicted Response Response 167.4 5 29.25 26 150 5 24.52 25. 10 2.5 5.5 20.25 2.42 Figure 9 shows the mathematical adjust of those values in Table 5, which generated a R 2 = 0.9195, revealing a linear mathematical relation among them. Thus, it is proved that the regression model is significant. predicted value 26.5 26 25.5 25 24.5 24 2.5 19.5592 16.5627 1.566 6.00 5.50 C: ph 4.50 Figure 9. Parity plot on predicted and actual values for validation study predicted vs actual R 2 = 0.9195 2 0 5 10 15 20 25 0 5 actual vaue IV. CONCLUSION The present study has been performed to produce by two stage conversion process namely; hydrolysis and fermentation using sago starch as a substrate. Sago starch was supplied from Sarawak Craun Research Institute. The production of bio from sago starch in this project has proven to be feasible in somewhat its large scale production promising high economic advantages as well as fulfill the demand for. Likewise, this laboratory scale study also offers the basis in scaling up the potential process for the production of bio in industrial scale. The optimum % (DE) has been obtained at was 22.76% (g glucose/g starch) with optimal parameters at 75.87 U/g glucoamylase, 28.49% substrate concentration and 2 hours incubation. The results obtained from validation study proved that the optimize parameters can be used for the improvement of the process. However, this value is comparatively low with other studies and further improvement throughout the hydrolysis stage can be carried out to overall improve the production. ACKNOWLEDGMENT We would like to acknowledge the Faculty of Engineering IIUM Malaysia for the support. Also, we thank Najiah Nadir for her useful guidance, Mohd Hafizul Shaibon and Sukiman Sengat for their technical assistance. REFERENCES [1] Senchez, O.J. and Cardona, C.A. (2007). Trends in biotechnological production of fuel from different feedstocks. Biosource Technology, 99: 5270-5295. doi 10.1016/j.biortech.2007.11.01 [2] Balat, M., Balat, H., and Öz, C. (2008). Progress in bio processing. Prograss in Energy and Combustion Science, 4: 551-57. [] Chiaramonti, D. (2007). Improvement of Crops Plants for Industrial End Uses. Netherlands: Springer. doi 10.1007/978-1-4020-5486-0-8. [4] Singhal, R. S., Kennedy, J.F., Gopalakrishnan SM, Kaczmarek A, Knill CJ, and Akmar P.F. (2008). Industrial production, processing, and utilization of sago palm-derived products. Carbohydrate Polymers, 72: 1-20. doi 10.1016/j.carbpol.2007.07.4 [5] Karim, A.A. Tie, A.P., Manan, D.M.A and Zaidul, I.S.M. (2008). Starch from Sago (Metroxylon sagu) Palm Tree: Properties, Prospects, and Challenges as a New Industrial Source for Food and Other Uses. Comprehensive Reviews in Food Science and Safety, 7: 215-228. [6] Ratnam, B. V. V., Rao, M. N., Rao, M. D., Rao, S. S., and Ayyanna, C. (200). Optimization of fermentation conditions for production of from sago starch using surface response methodology. World Journal of Microbiology and Biotechnology. Netherlands: Kluwer Academic Publisher. 19:52-526. [7] Abd-Aziz, S., Ang, D. C., Yusof, H. M., Karim, M. I. A., Ariff, A. B., and Uchiyama, K. (2001). Effect of C/N ratio and starch concentration on production from sago starch using recombinant yeast. World Journal of
Microbiology and Biotechnology. Netherlands: Kluwer Academic Publisher. 17:71-719. [8] Polycarpou, P. (2009). Bio production from Asphodelus aestivus. Renewable Energy, xxx:1-. [9] Mojović L., Nikolić S., Rakin M. and Vukašinović M. (2006). Production of bio fromcorn meal hydrolyzates. Fuel, 85:1750-1755. [10]Hezell, P., and Pachauri, R.K. (2006). Bioenergy and Agriculture: Promises and Challenges. Focus, 4:1-12. [11]Wang, W.J., Powell, W.J., and Oates, C.G. (1995). Pattern of enzyme hydrolysis in raw sago starch: effects of processing history. Carbohydrate Polymer, 26: 91-97. [12]Nadir, N., Mel, M., Karim, M.I.A., and Yunus, R.M. (2009). Comparison of Sweet Sorghum and Cassava for Ethanol Production by using Sacchromyces cerevesiae. Journal of Applied Science. 9(17):068-07. [1]Miller, G.L. (1959). Use of dinitrosalicylic acid reagent for determintation of reducing sugar. Anal. Chem. 1:426-428. [14]Shariffa, Y.N., Karim, A.A., Fazilah, A., and Zaidul, I.S.M. (2009). Enzymatic hydrolysis of granular native starches at sub-gelatinization temperature. Food Hydrocolloids, 2:44-440. [15] Bandaru, V.V.R., Somalanka, S.R., Mendu, D.R., Madicherla, N.R., Chityala, A. (2006). Optimization of fermentation conditions for production of ethnanol from starch by co-immobilized amyloglucosidase and cells Z. mobilis using response surface methodology. Enzyme and Microbial Technology, 8:209-214. [16]Tanyildizi, M.S., Dursun, O., and Murat, E. (2005). Optimization of α-amylase production by Bacillus sp. using surface response methodology. Process Biochemistry,:8-9. [17] Muralidhar, R.V., Chirumamila, R.R., Marchant, R. (2001). A response surface approach for comparison of lipase production by Candida cylindracea using two different carbon sources. Biochemical Engineering Journal, 9:17-2.