Scholarly article on topic 'Biodiesel production from Sesamum indicum L. seed oil: An optimization study'

Biodiesel production from Sesamum indicum L. seed oil: An optimization study Academic research paper on "Chemical sciences"

CC BY-NC-ND
0
0
Share paper
Academic journal
Egyptian Journal of Petroleum
OECD Field of science
Keywords
{" Sesamum indicum " / Biodiesel / Transesterification / Optimization / CCD / "Fuel properties"}

Abstract of research paper on Chemical sciences, author of scientific article — F.A. Dawodu, O.O. Ayodele, T. Bolanle-Ojo

Abstract Transesterification of Sesamum indicum L. oil was carried with methanol in the presence of sodium methoxide and the parameters affecting the reaction; vegetable oil/methanol molar ratio, catalyst concentration, reaction temperature and time were fully optimized by employing Central Composite Design method (CCD). A quadratic polynomial was developed to predict the response as a function of independent variables and their interactions and only the significant factors affecting the yield were fitted to a second-order response surface reduced 2FI model. At the optimum condition of 1:6 oil/methanol molar ratio, catalyst concentration of 0.75% and reaction time of 30min, biodiesel yield of 87.80% was achieved. The selected fuel properties were within the range set by ASTM and EN bodies.

Academic research paper on topic "Biodiesel production from Sesamum indicum L. seed oil: An optimization study"

Egyptian Journal of Petroleum (2014) xxx, xxx-xxx

Egyptian Petroleum Research Institute Egyptian Journal of Petroleum

www.elsevier.com/locate/egyjp www.sciencedirect.com

FULL LENGTH ARTICLE

Biodiesel production from Sesamum indicum L. seed oil: An optimization study

F.A. Dawodu a, O.O. Ayodele ab c *, T. Bolanle-Ojo c

a Department of Chemistry, University of Ibadan, Oyo State, Nigeria

b Key Laboratory of Green Process and Engineering, Institute of Process Engineering, CAS, Beijing, China c Department of Forest Products Development and Utilization, Forestry Research Institute of Nigeria, Oyo State, Nigeria

Received 9 June 2013; accepted 16 July 2013

KEYWORDS

Sesamum indicum; Biodiesel;

Transesterification;

Optimization;

Fuel properties

Abstract Transesterification of Sesamum indicum L. oil was carried with methanol in the presence of sodium methoxide and the parameters affecting the reaction; vegetable oil/methanol molar ratio, catalyst concentration, reaction temperature and time were fully optimized by employing Central Composite Design method (CCD). A quadratic polynomial was developed to predict the response as a function of independent variables and their interactions and only the significant factors affecting the yield were fitted to a second-order response surface reduced 2FI model. At the optimum condition of 1:6 oil/methanol molar ratio, catalyst concentration of 0.75% and reaction time of 30 min, biodiesel yield of 87.80% was achieved. The selected fuel properties were within the range set by ASTM and EN bodies.

© 2014 Production and hosting by Elsevier B.V. on behalf of Egyptian Petroleum Research Institute.

1. Introduction

The demand for energy is increasing at a substantial rate as the economy of the populous developing countries is growing. Currently, this high energy demand mainly depends on fossil fuel resources [1] but it is unsustainable and its exploitation leads to environmental degradation and increased emission

* Corresponding author at: Department of Forest Products Development and Utilization, Forestry Research Institute of Nigeria, Oyo State, Nigeria. Tel.: +234 803 3535 437. E-mail address: bayodele2002@gmail.com (O.O. Ayodele). Peer review under responsibility of Egyptian Petroleum Research Institute.

^jjfl I

Elsevier I Production and hosting by Elsevier

of greenhouse gases. Hence, the use of alternative sources of energy, such as biofuels, is attracting the interest of researchers [2].

In recent years, biodiesel has gained international attention as a source of alternative fuel due to characteristics like high degradability, low toxicity and emission of carbon monoxide, particulate matter and unburned hydrocarbons [3,4]. Biodiesel is a mixture of alkyl esters and it can be used in conventional compression ignition engines, which need almost no modification. Biodiesel can be used as heating oil and also as fuel [5,6]. So far, this alternative fuel has been successfully produced by transesterification of vegetable oils and animal fats using homogeneous basic catalysts.

Currently, partially or fully refined and edible-grade vegetable oils, such as soybean, rapeseed and sunflower, are the predominant feedstock for biodiesel production [7,8], which obviously results in the high price of biodiesel. Therefore, exploring ways to reduce the cost of raw material is of much

1110-0621 © 2014 Production and hosting by Elsevier B.V. on behalf of Egyptian Petroleum Research Institute. http://dx.doi.org/10.1016/j.ejpe.2014.05.006

interest in recent biodiesel research. As a result, in some countries, non-edible oils such as Jatropha curcas or waste cooking oils [9-11] are preferred due to their low price.

Realistically, non-edible oils only cannot meet the demand of energy consumption therefore, it has to be supplemented from some edible oils. For example, the average US production of soybean oil from 1993 to 1995 was 6.8 billion kg, and in 2002, soybeans were harvested from more than 30 million hectares across the United States, which accounts for 40% of the total world soybean output [12]. This production capacity accounts for more than 50% of the total available bio-based oil for industrial applications. Also rapeseed is mostly grown in Europe, China and India and there are many published reports on the utilization of these seed oils as biodiesel fuels [13-15].

Beniseed (Sesamum indicum L), herbaceous plant in Nigeria as well as in India, China, Sudan, Burma, Bangladesh, Indonesia, Egypt, Tunisia, belongs to the family of Pedaliaceae. Sesame seed has one of the highest oil contents of any seed and is considered to be the oldest oilseed crop known to man, highly resistant to drought and has the ability of growing where most crops fail [16,17]. The seed colour varies from cream-white to charcoal black but it is mainly black. In Nigeria, the notable colours of sesame seed are white, yellow and black [18]. The major world producers include India, Sudan, China and Burma while Nigeria and Ethiopia are also major producers and exporters [19].

The fat of sesame is of importance in the food industry due to its flavour and stability, its oil has been found to contain sesamin and sesaminol lignans in its non-glycerol fraction, which are known to play an important role in the oxidative stability and antioxidant activity [20]. The main fatty acid composition of the oil is oleic (43%), linoleic (35%), palmitic (11%) and stearic (7%) acids, contributing about 96% of the total fatty acids [17,21]. It also contains some polyunsaturated fatty acids basically Omega 6 fatty acids but lacks Omega 3 fatty acids.

Sesame seed is cultivated and produced in large quantity in Nigeria, especially in the Northern part of the country and is under-utilized in some parts of the country. Therefore, there is a greater need to utilize some for energy purposes. Based on the recent statement that Nigeria envisions an energy transition from crude oil to renewable energy [22], we therefore investigated the oil of sesame seed as an alternative feedstock for the production of biodiesel fuel.

In this work, we produced our biodiesel from sesame oil through transesterification reaction in the presence of an alkali based catalyst and the factors affecting the reaction were fully optimized by following the factorial design and response surface methodology [23]. Limited reports on biodiesel production from the oil of S. indicum and its optimization using Central Composite Design technique exist in the literature [18,24].

2. Materials and methods

2.1. Materials

Seeds of S. indicum L. were bought from the open market, dried to an acceptable moisture level and milled with a laboratory milling machine. The oil-seed was then extracted with n-hexane using a soxhlet apparatus and characterized according to the AOCS official methods [25] (Table 1). Analytical grades of sul-

Table 1 Central Composite Design for transesterification reaction.

Variables (coded factors) Levels

-1 0 + 1

Molar ratio of oil to methanol (Xj) 6:1 1:9 1:12

Catalyst/oil ratio (X2) 0.75 1.00 1.25

Reaction temperature (X3) 50 60 70

Reaction time (X4) 30 60 90

phuric acid, methanol (Beijing Chemical works), sodium hydroxide, and hexane (Xilong Chemicals) were used without any further purification. The reference standard of fatty acid methyl esters (methyl palmitate, methyl stearate, methyl linole-ate and methyl linolenate) was purchased from accustandard, methyl oleate (J&K Chemicals) and monolein from the Tokyo Chemical Industry, while diolein and triolein (Sinopharm Chemicals) and glycerol standards were purchased from Xilong Chemicals Co.

2.2. Methods

2.2.1. Experimental procedure

Reactions were carried out in a 250 cm3 two-necked batch reactor. The reactor was initially filled with the desired amount of oil, and then placed in the constant-temperature oil bath equipped with reflux condenser, stopper and heated to a predetermined temperature. The catalyst, sodium methoxide was generated by dissolving anhydrous sodium hydroxide in methanol and the resulting solution was added to the agitated reactor. At the

Table 2 Physical and chemical parameters of Sesamum indicum oil.

Parameters Values

% Yield 61.99 ± 0.37

Colour Light yellow

State at room temperature Liquid

Specific gravity (25 °C) 0.8525 ± 0.03

Viscosity at 40 °C (mm2/s) 22.63

Acid value (mgKOH/g) 3.15 ± 0.58

FFA (%) 1.58 ± 0.29

Saponification value (mgKOH/g) 142.2 ± 2.40

Iodine value (mg I/g) 86.15 ± 1.63

Peroxide value (mgO2/g) 2.8 ± 0.00

Table 3 Fatty acid composition of Sesamum indicum oil.

Fatty acid % Composition

Caproic acid (6:0) 7.80

Palmitic acid (16:0) 6.80

Stearic acid (18:0) 8.98

Oleic acid (18:1) 28.54

Linoleic acid (18:2) 39.73

Linolenic acid (18:3) 0.31

Lignoceric acid (24:0) 4.59

Others 3.25

Table 4 Result of CCD for transesterification of Sesamum indicum oil.

Run A B C D X1 X2 X3 X4 % yield

1 6:1 0.75 50 30 -1 -1 -1 -1 87.80

2 12:1 0.75 50 30 + 1 -1 -1 -1 77.44

3 6:1 1.25 50 30 -1 +1 -1 -1 65.68

4 12:1 1.25 50 30 + 1 +1 -1 -1 80.40

5 6:1 0.75 70 30 -1 -1 +1 +1 75.40

6 12:1 0.75 70 30 + 1 -1 +1 -1 76.36

7 6:1 1.25 70 30 -1 +1 +1 -1 78.12

8 12:1 1.25 70 30 +1 +1 +1 -1 81.48

9 6:1 0.75 50 90 -1 -1 -1 +1 82.88

10 12:1 0.75 50 90 +1 -1 -1 +1 78.08

11 6:1 1.25 50 90 -1 +1 -1 +1 75.68

12 12:1 1.25 50 90 +1 +1 -1 -1 79.68

13 6:1 0.75 70 90 -1 -1 +1 +1 81.00

14 12:1 0.75 70 90 +1 -1 +1 +1 81.20

15 6:1 1.25 70 90 -1 +1 +1 +1 80.00

16 12:1 1.25 70 90 +1 +1 +1 +1 86.24

17 9:1 1.00 60 60 0 0 0 0 78.68

18 9:1 1.00 60 60 0 0 0 0 77.52

19 9:1 1.00 60 60 0 0 0 0 74.86

20 9:1 1.00 60 60 0 0 0 0 78.90

21 9:1 1.00 60 60 0 0 0 0 69.41

22 9:1 1.00 60 60 0 0 0 0 75.77

23 9:1 1.00 60 60 0 0 0 0 77.21

24 9:1 1.00 60 60 0 0 0 0 73.87

A-Methanol:seed oil molar ratio; B-Catalyst/seed oil (%wt); C-Temperature (oC); D-Time (min); Xj, X2, X3, and X4 are coded factor for A, B, C, and D respectively; Rate of mixing is constant (800 rpm).

end of the reaction, excess methanol was removed using a rotary evaporator and the product transferred to a separating funnel for phase separation. The lower phase rich in glycerol was separated and the upper phase was then washed with lukewarm distilled water. The percentage ester yield was calculated by comparing the weight of methyl ester with the initial weight of the oil (%w/w). Biodiesel purity, that is, the methyl ester concentration (%wt) in the product was calculated by Gel Permeation Chromatography (GPC) (Asahipak GF310-HQ column, oven temperature: 40 0C) with a refractive index as detector. Acetone was used as the mobile phase at flow rate 1 ml/min. This also allows for the quantification of the monoglyceride, diglyceride and triglyceride contents in the biodiesel [26].

2.3. Experimental design

Central Composite Design technique (CCD) was used for optimization of biodiesel production. The experimental design applied to this work was a full 24 factorial design (four factors each at two levels) with 8 centre points making a total of 24 runs (Table 1). Experiments were run at random to minimize errors due to possible systematic trends in the variables.

Minitab 16.0 was used for normal plot and pareto chart of the standardized effects. The effects allowed for full investigation of the significant parameters affecting transesterification reaction. The Design Expert 6.0 software was therefore used for regression and graphical analyses of the data obtained. Statistical analysis

Factor Name

A Methanol: oil

B Catalyst weight

C Temperature

D Time

0.0 0.5 1.0 l.S 2.0 2.5 3.0 3.5 Standardized Effect

Figure 1 Pareto chart showing the effects of different variables on FAME yield.

Effect Type • Not Significant ■ Significant

Factor Name

A Methanol: oil

B Catalyst weight

C Temperature

D Time

Figure 2 Normal plot of the standardized effects (response is %Yield, Alpha = 0.05).

Table 5 Results of ANOVA for response surface reduced 2FI model.

Significant

Source Sum of square DF Mean square F-value P-value

Model 228.27 5 45.65 2.96 0.0404

A 12.82 1 12.82 0.83 0.3744

B 10.37 1 10.37 0.67 0.4233

C 9.24 1 9.24 0.60 0.4493

AB 111.94 1 111.94 7.25 0.0149

BC 83.91 1 83.91 5.43 0.0316

Residual 278.05 18 15.45

Lack of fit 106.71 3 35.57 3.11 0.0579

Pure error 171.34 15 11.42

Corr. Total 506.32 23

ss-not significant.

Predicted FAME >idd

Figure 3 Graph showing the relationship between residuals and predicted biodiesel yield.

of the model was performed to evaluate the analysis of variance (ANOVA). A quadratic polynomial was developed to predict the response as a function of independent variables and their interactions [27] and a second-order polynomial equation [28] was used.

2.4. Fuel properties

The fatty acid methyl ester produced was subjected to fuel properties tests such as density, kinematic viscosity, and acid number.

Y = b0 + £ biXi + £ bjX] + E YjbjXXj

3. Results and discussion

3.1. Characterization of S. indicum oil

where I and j are the linear and quadratic coefficients respectively, b is the regression coefficient, k is the number of factors studied and optimized and e is the random error.

The physical and chemical characterization of the oil was determined according to a standard procedure (Table 2) and was compared to the results found in the literature [29]. The %FFA was within the limit allowed for alkali-catalysed transesterification

Figure 4 Predicted and actual experimental results for the model.

Residuals

Figure 5 Normal plot of residuals.

-1.00 -1.00

Figure 6 Three dimensional plot of the effect of oil/methanol molar ratio and catalyst concentration on the yield of FAME.

reaction [30]. The fatty acid composition was determined by gas chromatography and is shown in Table 3. Oleic acid and linoleic acid were predominant in the oil.

3.2. Experimental design

In this study, optimization of biodiesel yield was carried out using Central Composite Design to fully establish parameters affecting transesterification reaction. Table 4 summarizes the yield of FAME from S. indicum oil.

Based on a = 0.05, only factors with P-value less than 0.05 are significantly affecting % yield as observed in Fig. 1. The significant factors are further confirmed by normal plot of the standardized effect (Fig. 2). Only the terms affecting the yield were then included in the result of the second-order response surface reduced 2FI model fitting in the form of ANOVA shown in Table 5. The Fisher F-test (F = 2.96) with a low P-value less than 0.05 indicates a high significance for the model. There is only a 4.04% chance that a ''model F-value'' of 3.84 could be due to noise factors in the experiments. With low P-value from the analysis of variance, the reduced qua-

-1.00 -1.00

Figure 7 Three dimensional plot of the effect of oil/methanol molar ratio and catalyst concentration on the yield of FAME.

-loo о.ьо ооо о.я 1 ж

catalyst wt%

Figure 8 Contour plot of the interaction between oil/methanol molar ratio and catalyst weight.

dratic polynomial model was significant and sufficiently summarizes the relationships between the response and the significant variables. Therefore, the terms AB and BC are significant model terms and were fitted into a second order reduced 2FI model in term of coded factors as shown Eq. (2). The Lack of Fit F-value'' of 3.11 implies the Lack of Fit is not significant relative to the pure error. Therefore, the model proposed is valid and statistically significant.

Y = 78.07 + 0.89A - 0.80B + 0.76C + 2.64AB + 2.29BC (2)

where Y is the biodiesel yield and A, B, and C are coded factors for vegetable oil/methanol ratio, catalyst weight and temperature respectively.

The model predicts that yield of 78.07% could be obtained under the optimum operating conditions and catalyst weight has the highest interaction between all the variables considered in this experiment.

To test the fit of the model, the residual distribution graph was plotted and it is observed that the residual distribution does not follow a particular trend with respect to the predicted response values. This tests the assumption of constant variance and should be a random scatter as seen in Fig. 3. The residuals are less than 3% which indicates the model accuracy on the influence of FAME yield over the experimental factors studied. Fig. 4 shows the relationships between the experimental response values and predicted response values. This graph helps to detect a value, or group of values, that are not easily predicted by the

temperature (°C)

Figure 9 Contour plot of the interaction between catalyst weight and reaction temperature.

Figure 10 Chromatogram of FAME at the optimum condition before washing (a) unreacted triglyceride; (b) FAME; (c) glycerol; (d) impurities.

model. To test the fit of the model, the data points should be split evenly by the 45 degree line. Fig. 4 agrees with this assumption and the fit of the model was further confirmed by the normal plot of residuals (Fig. 5) which shows the linearity of the residuals. This indicates the errors are normally distributed for all the reduced responses. It can be therefore concluded that the reduced 2FI model adequately correlates the relationship between the reaction variables and the FAME yield.

3.3. Interaction and responses

The statistical analysis signifies catalyst weight as the most important interacting factor in the biodiesel yield response.

Catalysts are known to speed up a chemical reaction and could also inhibit or slow down a reaction if they are in limited quantity or in excess. As seen in Eq. (2), catalyst weight has a negative effect on the reaction but its interaction with other factors; methanol to vegetable oil molar ratio and temperature, show positive interactions. Therefore, the biodiesel yield increases when the values of these factors increase. The meth-anol/oil molar ratio interaction is very significant according to the statistical analysis and shows a positive influence on the response, biodiesel yield. Likewise, temperature has a positive influence on the reaction according to Eq. (2).

In order to fully understand the interaction between the significant variables, three dimensional surface response plots

wavelength (cm )

Figure 11 Infrared spectrum of FAME from Sesamum indicum L.

Table 6 Characteristic infrared bands of biodiesel.

Wave number Present Group Vibration

(cm work assignment type

720 723.09 -CH2 Plane rocking

1112- 1070 1098.48 -C-O Stretching

1300- 1100 1168.32 -C-O Stretching

1275- 1100 1241.51 -CH In-plane bending

1475- 1350 1377.33 -CH2, -CH3 Bending

1472- -1427 1464.89 -CH3 Asymmetric bending

1750- 1730 1744.11 -C=O Stretching

3000- -2800 2854.28 -CH2 Symmetric stretching

3000- -2800 2925.90 -CH2 Asymmetric stretching

3050- 3000 3007.92 -CH Stretching

were employed. When twice the amount of stoichometric ratio was used with catalyst ratio of 0.75%, the yield of 87.80% was achieved in 30 min (Fig. 6). The yield drastically reduced when the catalyst ratio is increased to 1.25%. We observed an emulsion (soap) in the reaction system and this significantly reduced the yield of our product. An improved yield was achieved when the oil/methanol ratio was increased keeping the catalyst ratio at 1.25%. This helps in further diluting the reaction system and soap formation was not observed. Even little amount of FFA in refined vegetable oil could still form an emulsion if excess sodium or potassium hydroxide is used. Fig. 7 shows the relationship between catalyst weight and temperature. Overall, catalyst plays an important role in the reaction and this is further confirmed with the contour plots (Figs. 8 and 9).

3.4. Analysis of FAME from S. indicum L. oil

FAME produced at the optimum operating condition was analysed using the GPC method as seen in Fig. 10. There was high conversion of triglyceride to biodiesel at 1:6 M ratio of oil to methanol, catalyst concentration of 0.75 wt%, temperature of 50 0C and reaction time of 30 min. Conversion of TG to biodiesel was further confirmed based on the functional groups of the product (Fig. 11). The two most polar bonds in esters (containing the —CO—O—C— unit) are the C=O and C—O bonds and these bonds produce the strongest bands in the spectrum of any ester (summarized in Table 6). Aliphatic esters produce C=O and C—O bands at 1750-1730 and 1300-1100 cm-1 respectively, an indicator of conversion of triglyceride to FAME. Acid number after production (0.36 mgKOH/g), acid number after six months of storage (3.73 mgKOH/g), density (0.8809 g/mL at 15 0C), kinematic viscosity (4.9 mm2/s at 40 0Q are within the values specified by international standards.

4. Conclusion

Transesterification of S. indicum oil to biodiesel was fully optimized by Central Composite Design Technique and the optimum condition was found at oil/methanol molar ratio of 1:6, catalyst weight of 0.75 wt% relative to the weight of oil used, temperature of 50 0C and reaction time of 30 min. The purity of the biodiesel, as detected via GPC, was high and the properties were within the specified limit. Although, the acid number increased significantly after six months of storage due to unsaturation of the fatty acid chains, this effect could be stabilized with some recommended antioxidants. Therefore, sesame seed

oil could serve as a potential feedstock for biodiesel production and also complement the existing biodiesel feedstocks.

Acknowledgements

The World Academy of Sciences (TWAS) and the Chinese Academy of Sciences (CAS) are gratefully acknowledged.

References

[1] K. Abebe, Y.K. Endalew, R. Zanzi, Biomass Bioenergy 35 (2011) 3709-3787.

[2] K. Cheenkachorn, B. Fungtammasan, Int. J. Green Energy 6 (2009) 57-72.

[3] G. Vicente, A. Coteron, M. Martinez, J. Aracil, Ind. Crop. Prod. 8 (1998) 29-35.

[4] S. Al Zuhair, Biofpr 1 (2007) 57-66.

[5] G. Mushrush, E.J. Beal, G. Spencer, J.H. Wynne, C.L. Lloyd, J.M. Hughes, et al, J. Environ. Sci. Health A 36 (2001) 613-622.

[6] D.A. Wardle, Renew. Sust. Energy Rev. 7 (2003) 1-64.

[7] M.J. Haas, Fuel Process Technol. 86 (2005) 1087-1096.

[8] X. Fan, R. Burton, G. Austic, Int. J. Green Energy 7 (2010) 593599.

[9] Y. Wang, O. Shiyi, L. Pengzhan, Z. Zhang, Energy Convers. Manage. 48 (2007) 184-188.

[10] G.L.N. Rao, B.D. Prasad, S. Sampath, K. Rajagopal, Int. J. Green Energy 4 (2007) 645-658.

[11] T.M. Labib, S.I. Hawash, K.M. El-Khatib, A.M. Sharaky, G.I. El Diwani, E. Abdel Kader, Egyptian J. Petrol. 22 (2013) 9-16.

[12] A.J. Kinney, T.E. Clemente, Fuel Process. Technol. 86 (2005) 1137-1147.

[13] G. Vicente, M. Martinez, J. Aracil, Ing. Quim. 31 (1999) 153-159.

[14] F. Qiu, Y. Li, D. Yang, X. Li, P. Sun, Appl. Energy 88 (2011) 2050-2055.

[15] B.K. Sylvain-Didier, J. Fuel Chem. Technol. 39 (2011) 258-264.

[16] H. Koca, M. Bor, F. Ozdemir, I. Turkan, Environ. Exp. Bot. 60 (2007) 344-351.

[17] M. Elleuch, S. Besbes, O. Roiseux, C. Blecker, H. Attia, Food Chem. 103 (2007) 641-650.

[18] S. Fariku, A.E. Ndonya, P.Y. Bitrus, Afr. J. Biotech. 6 (2007) 2442-2443.

[19] M.K.S. El Khier, K.E.A. Ishag, A.E.A. Yagoub, Res. J. Agric. Biol. Sci. 4 (2008) 761-766.

[20] W.H. Wu, Food Chem. 104 (2007) 341-344.

[21] C. Borchani, S. Besbes, C.H. Blecker, H. Attia, J. Agric. Sci. Technol. 12 (2010) 585-596.

[22] Energy Commission of Nigeria (ECN), Renewable Energy Master Plan; Executive Summary, 2005.

[23] A.K. Tiwari, A. Kumar, H. Raheman, Biomass Bioenergy 31 (2007) 569-575.

[24] S. Abdurrahman, M. Zahir, K. Duz, B.K. Canan, H. Aylin, Canada Bioresour. Technol. 99 (2008) 6656-6660.

[25] AOCS, Official Methods and Recommended Practices of the American Oil Chemists' Society, AOCS Press (1998), Champaign, IL.

[26] D. Darnoko, M. Cheryan, E.G. Perkins, J. Liq. Chromatogr. Rel. Technol. 23 (2000) 2327-2335.

[27] D.C. Montgomery, Design and analysis of experiments, John Wiley and Sons, New York, 2001.

[28] R. Balusu, R R. Paduru, S.K. Kuravi, G. Seenayya, G. Reddy, Process Biochem. 40 (2005) 3025-3030.

[29] F.A. Dawodu, EJEAFChe 8 (2009) 102-110.

[30] M P. Dorado, E. Ballesteros, J.A. Almeida, C. Schellet, H P. Lohrlein, R. Krause, Trans. ASAE 45 (2002) 525-529.