Reactive Extraction of Jatrophacurcas L. Seed for Production ofBiodiesel: Process OptimizationStudyS I E W H O O N G S H U I T , †

K E A T T E O N G L E E , * , †

A Z L I N A H A R U N K A M A R U D D I N , † A N DS U Z A N A Y U S U P ‡

School of Chemical Engineering, Engineering Campus,Universiti Sains Malaysia, Seri Ampangan,14300 Nibong Tebal, Pulau Pinang, Malaysia, andDepartment of Chemical Engineering, Universiti TeknologiPETRONAS, 31750 Tronoh, Perak, Malaysia

Received September 8, 2009. Revised manuscript receivedApril 20, 2010. Accepted April 27, 2010.

Biodiesel from Jatropha curcas L. seed is conventionallyproduced via a two-step method: extraction of oil and subsequentesterification/transesterification to fatty acid methyl esters(FAME), commonly known as biodiesel. Contrarily, in this study,a single step in situ extraction, esterification and transesteri-fication (collectively known as reactive extraction) of J. curcasL. seed to biodiesel, was investigated and optimized. Designof experiments (DOE) was used to study the effect of variousprocessparametersontheyieldofFAME.Theprocessparametersstudied include reaction temperature (30-60 °C), methanolto seed ratio (5-20 mL/g), catalyst loading (5-30 wt %), andreaction time (1-24 h). The optimum reaction condition was thenobtained by using response surface methodology (RSM)coupled with central composite design (CCD). Results showedthat an optimum biodiesel yield of 98.1% can be obtainedunder the following reaction conditions: reaction temperatureof 60 °C, methanol to seed ratio of 10.5 mL/g, 21.8 wt % of H2SO4,and reaction period of 10 h.

IntroductionOwing to the world petroleum crisis, production of biodieselin Malaysia has increased steadily for the past three years;from 325,000 tons in 2006 to 400,000 tons in 2007 and 420,000tons in 2008 (1). This increase is in tandem with the world’sincreasing biodiesel demand, especially in the Europeanregion. However, the practice of using edible oils, which iscurrently the most common feedstock for biodiesel produc-tion, has raised criticism from various sectors, especiallynongovernmental organization (NGO), claiming that biodie-sel is competing for resources with the food industry.Therefore, production of biodiesel from nonedible oils suchas Jatropha curcas L. seeds (2), beef tallow (3), waste cookingoil (4), and Cerbera odollam (sea mango) (5) would be a

potential solution to this issue. Among these choices, J. curcasL. seed has recently been hailed as the promising feedstockfor biodiesel production because it can be cultivated in dryand marginal lands (6), and thus it does not compete forarable land that would have otherwise being planted withfood crops. Besides, its oil yield as shown in Table 1 iscomparable to that of palm oil but much higher as comparedto other edible oil crops such as rapeseed, sunflower, andsoybean (7, 8). Therefore, even Malaysia, the world’s secondlargest palm oil producer, is now diversifying its biodieselfeedstock toward jatropha. The total jatropha plantation areain Malaysia, at the end of 2008, was estimated at 750,000acres and is expected to increase to 1.5 million acres in 2009and 2.5 million acres in 2010 (9).

Conventional methods for producing biodiesel fromjatropha and other types of oil seeds involve various stages:oil extraction, purification (degumming, dewaxing, deacidi-fication, dephosphorization, dehydration, etc.), and subse-quent esterification or transesterification. These multiplebiodiesel processing stages constitute >70% of the totalbiodiesel production cost if refined oil is used as feedstock(10). Recently, in our previous study, it was shown that insitu extraction and esterification/transesterification, simplyknown as reactive extraction, is a feasible technology for theproduction of biodiesel using a single step that can cut theprocessing cost. In the reactive extraction process, extractionof oil and esterification/transesterification proceed in a singlestep in which the oil-bearing material contacts with alcoholdirectly instead of reacting with pre-extracted oil. In otherwords, alcohol acts both as an extraction solvent and as atransesterification reagent during reactive extraction, andtherefore a higher amount of alcohol is required. However,reactive extraction eliminates the requirement of two separateprocesses, the costly hexane oil extraction process and thetransesterification reaction process, thus reducing processingtime, cost, and amount of solvent required (11). Furthermore,on the basis of a similar study reported in the literature (usingsoybeans), it was demonstrated that the reactive extractionprocess can be scaled up without encountering muchproblem in mass and heat transfer limitations (12).

Nevertheless, in our previous study, the requirement ofa 24 h reaction time to achieve a high fatty acid methyl esters(FAME) yield of 99.9% makes this process unattractive froman industrial perspective (11). Thus, the aim of this study isto optimize the process parameters of the acid-catalyzedreactive extraction process for the production of FAME fromJ. curcas L. seed.

Experimental SectionMaterials. J. curcas L. seed was purchased from Misi BumiAlam Sdn Bhd, Malaysia. Methanol (99.9% purity) waspurchased from J. T. Baker, Germany. The remainingchemicals used in this study, sulfuric acid (H2SO4, 95-97%purity), methyl heptadecanoate (internal standard), and puremethyl esters such as methyl palmitate, methyl stearate,

* Corresponding author phone:+604-5996467; fax:+604-5941013;e-mail: chktlee@eng.usm.my.

† School of Chemical Engineering, Engineering Campus, UniversitiSains Malaysia, Seri Ampangan, 14300 Nibong Tebal, Pulau Pinang,Malaysia

‡ Department of Chemical Engineering, Universiti TeknologiPETRONAS, 31750 Tronoh, Perak, Malaysia

TABLE 1. Oil Yield of Jatropha curcas L. Seed and OtherMajor Oil Crops

oil crop oil yield (tons/ha/year)

J. curcas L. seed 2.70 (7)oil palm (mesocarp) 3.62 (8)rapeseed 0.68 (8)sunflower 0.46 (8)soybean 0.40 (8)

Environ. Sci. Technol. 2010, 44, 4361–4367

10.1021/es902608v 2010 American Chemical Society VOL. 44, NO. 11, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 4361

Published on Web 05/10/2010

methyl oleate, and methyl linoleate, were purchased fromFluka Chemie, Germany.

Pretreatment of J. curcas L. Seed and Oil Content.Initially, fresh J. curcas L. seed was blended and sieved to asize of <1 mm (11). It was then weighed and dried in the ovenat 76 °C repeatedly until constant weight was achieved (13).The dried seed was then sieved again to obtain fine particlesof e0.355 mm in size. To determine the maximum amountof oil that can be extracted from the seed using theconventional method, a Soxhlet extractor with excess n-hexane as the solvent was utilized. After the extractionprocess, hexane was removed using a rotary evaporator, andthe extracted oil was measured (11). This value will be usedin the calculation of yield.

Reactive Extraction. The reactive extraction to convertjatropha seed to biodiesel was carried out in a round-bottomflask equipped with a reflux system, magnetic stirrer, andheater. Initially, 20 g of jatropha seed was loaded into a 500mL round-bottom flask. Methanol and concentrated H2SO4

in the ranges of 5-20 mL/g and 5-30 wt %, respectively,were then added into the round-bottom flask. After that, themixture was heated to the desired temperature (ranging from30 to 60 °C) for the necessary duration (ranging from 1 to24 h). Upon completion of the reaction period, the mixturewas cooled and then filtered. The solid residue was washedrepeatedly with methanol, and the excess methanol in thefiltrate was recovered using a rotary evaporator. Afterevaporation, two layers of liquid were formed. The upperlayer was dark yellow in color containing crude biodiesel,whereas the bottom layer was dark brown in color containingglycerol. The volume of the top layer was then measured andrecorded (11). Then, the upper layer was washed with 20%NaCl solution several times until the pH became neutral.After washing, the upper layer was dried over anhydroussodium sulfate.

Sample Analysis. The composition and yield of fatty acidmethyl esters (FAME) or biodiesel in the upper layer of thereactive extraction products were analyzed using gas chro-matography (PerkinElmer, claurus 500) equipped with a flameionized detector (FID) and a Nukol capillary column (15 m× 0.53 mm; 0.5 µm film). n-Hexane was used as the solvent,whereas helium was used as the carrier gas. The oventemperature was set at 110 °C and then increased to 220 °Cat a rate of 10 °C/min. The temperatures of the detector andinjector were set at 220 and 250 °C, respectively. Methylheptadecanoate was used as the internal standard. The peaksof different methyl esters were identified by comparing theretention time of each component in the reaction sampleswith the peak of pure methyl ester standard compounds.The yield of FAME in the samples was calculated as (11)

Statistical Analysis and Optimization Using Design ofExperiments (DOE). The effect of reactive extraction processparameters on the yield of FAME was studied using DOE.

The process parameters studied include reaction temper-ature, methanol to seed ratio, catalyst loading, and reactiontime. The DOE selected was response surface method (RSM)coupled with central composite design (CCD) using Design-Expert version 6.0.6 (Stat-Ease, Inc.) software. Table 2indicates the coded and actual values of the processparameters used, whereas Table 3 shows the complete designmatrix. On the basis of the statistical analysis result, the yieldof FAME was correlated to the process parameters with aquadratic model using regression analysis. It was then utilizedfor optimization purposes.

Results and DiscussionDevelopment of Regression Model. Table 3 shows the yieldof FAME obtained using reactive extraction with a combina-tion of various process parameters. The result was thenanalyzed using analysis of variance (ANOVA) and is shownin Table 4. By eliminating the insignificant parameters (valuesof “prob > F” of >0.05), multiple regression analysis gives thefollowing quadratic model equation (in coded factors) thatcorrelates the yield of FAME to the various processparameters:

As shown in Table 4, the F test (Fisher) on eq 2 gives anF value of 21.8 and a “prob > F” value of <0.0001, indicatingthat the developed model is significant. Besides that, R2 shownin Table 4 is very close to unity, 0.9532, indicating that thedeveloped model equation successfully captured the cor-relation between the process parameters to the yield of FAMEfor reactive extraction process of jatropha seed.

Effect of Single Process Parameter. On the basis of theresults shown in Table 4, all four process parameters studied,reaction time (A), reaction temperature (B), methanol to seedratio (C), and catalyst loading (D), were found to significantlyaffect the yield of FAME because their F values are higherthan the theoretical value, F1,15 of 4.54 at the 95% confidenceinterval. THe parameter with the highest F value will havethe most significant effect. Thus, by referring to Table 4, theparameter with the most significant effect on the yield ofFAME in descending order is catalyst (H2SO4) loading,followed by reaction temperature, reaction time, and finallymethanol to seed ratio. Apart from that, the positive sign forall four regression coefficients (A, B, C, and D) in eq 2 indicatesa positive effect on the yield of FAME. These findings can beeasily verified by visually inspecting the experimental resultsshown in Table 3. For example, by comparison between runs2 and 7 (and other comparable runs), an increase in H2SO4

loading caused a vast increase in the yield of FAME. A similarfinding can be made for the effect of reaction temperature(e.g., runs 6 and 22). However, for reaction time (runs 20 and25) and methanol to seed ratio (runs 3 and 13), increases inthese parameters caused only a slight increment in the yieldof FAME.

Figure 1 shows the effect of reaction time on the yield ofFAME. As shown in Figure 1, the yield of FAME increasedwith higher reaction time. In our previous study, it wasreported that the limiting factor for reactive extraction process

TABLE 2. Coded and Actual Values of Process Parameters in Central Composite Design

level

variable code unit -2 (-r) -1 0 +1 +2 (+r)

reaction time A h 1 7 13 18 24reaction temperature B °C 30 38 45 53 60methanol to seed ratio C g/mL 5 8.75 12.5 16.25 20H2SO4 loading D wt % 5 11.25 17.5 23.75 30

yield (%) )

(Σconcn of each component) × (vol of upper layer)total wt of oil in sample

× 100%

(1)

yield (wt %) ) 58.6 + 10.2A + 16.8B + 9.11C + 18.5D -5.34A2 - 4.28D2 + 5.63BD (2)

4362 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 11, 2010

is the leaching of oil from the seed itself (11). Therefore,sufficient time will be required to leach the oil from the seedand subsequently be transesterified to FAME. From Figure2, the yield of FAME was found to increase linearly withreaction temperature in agreement with those reported inthe literature (14, 15). This can be easily justified as highertemperature can increase the extraction rate of oil fromthe seed and, in addition, the endothermic nature of thetransesterification reaction will shift the reaction toward theforward direction at higher reaction temperature (16). It isnoted here that in this study, the reaction temperature islimited to 60 °C so that this process can be comparable withthe conventional base-catalyzed transesterification reaction.

The role of methanol in reactive extraction is not only asan extraction solvent but also as a transesterification reagent.Therefore, the requirement of methanol in this particularprocess is higher than in other conventional transesterifi-cation reaction. As shown in Figure 3, the yield of FAME wasfound to increase linearly with higher methanol to seed ratio.This is because with a higher amount of methanol, theconcentration gradient to leach oil from the seed will belarger. Furthermore, the excess amount of methanol will drivethe transesterification reaction forward toward the formationof FAME. However, it was found that the effect of methanolto seed ratio on the yield of FAME is least significant amongthe four process parameters studied as indicated by the

TABLE 3. Experiment Matrix with Coded Factors of CCD and Response

runA, reaction

time (h)B, reaction

temperature (°C)C, methanol to

seed ratio (g/mL)D, catalyst

loading (wt %) yield (%)

1 13 45 12.50 5.00 4.232 7 38 16.25 11.25 19.203 18 53 16.25 11.25 60.764 13 45 5.00 17.50 30.205 7 53 8.75 11.25 16.826 13 60 12.50 17.50 99.547 7 38 16.25 23.75 48.588 7 38 8.75 23.75 29.659 7 38 8.75 11.25 1.7510 7 53 16.25 23.75 86.2411 18 53 8.75 23.75 90.9312 18 53 16.25 23.75 94.7813 18 53 8.75 11.25 52.2714 18 38 8.75 23.75 40.6815 18 38 8.75 11.25 22.2216 1 45 12.50 17.50 15.5417 7 53 8.75 23.75 70.1518 13 45 20.00 17.50 93.1919 7 53 16.25 11.25 28.7520 24 45 12.50 17.50 68.4021 13 45 12.50 30.00 88.1722 13 30 12.50 17.50 28.0023 18 38 16.25 11.25 30.6524 18 38 16.25 23.75 48.29repeated experiments

25 13 45 12.50 17.50 60.8326 13 45 12.50 17.50 60.4527 13 45 12.50 17.50 56.6628 13 45 12.50 17.50 60.3929 13 45 12.50 17.50 53.2330 13 45 12.50 17.50 60.13

TABLE 4. ANOVA for Response Surface Quadratic Model for the Yield of FAME

source sum of squares DF mean square F value prob > F

model 21688.99 14 1549.21 21.84 <0.0001a

A 2504.31 1 2504.31 35.30 <0.0001a

B 6758.98 1 6758.98 95.27 <0.0001a

C 1994.00 1 1994.00 28.11 <0.0001a

D 8242.14 1 8242.14 116.18 <0.0001a

A2 788.11 1 788.11 11.01 0.0047a

B2 0.35 1 0.35 0.004984 0.9447b

C2 4.5 1 4.5 0.063 0.8045b

D2 502.2 1 502.2 7.08 0.0178a

AB 183.06 1 183.06 2.58 0.1290b

AC 81.09 1 81.09 1.14 0.3019b

AD 219.93 1 219.93 3.10 0.0987b

BC 9.09 1 9.09 0.13 0.7254b

BD 507.6 1 507.6 7.15 0.0173a

CD 0.002025 1 0.002025 0.00002854 0.9958b

residual 1064.19 15 70.95R2 ) 0.9532; adjusted R2 ) 0.9096; predicted R2 ) 0.7394; standard deviation ) 8.42; mean ) 50.69a Significant at 95% confident interval. b Not significant at 95% confident interval.

VOL. 44, NO. 11, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 4363

smallest F value shown in Table 4. This is probably becauseeven at the lower range of methanol to seed ratio used in thisstudy (5), there is already sufficient methanol to create theconcentration gradient required to leach oil from the seed.

As mentioned in the earlier section, catalyst (H2SO4)loading was found to have the most significant positive effecton the yield of FAME among all four reaction parameters.From Figure 4, it can be seen that a higher amount of catalyst

used will increase the yield of FAME significantly. Forinstance, when the amount of H2SO4 in the reaction mixturewas set at a minimum level of 5 wt % (run 1), the yield ofFAME was merely 4.23%; however, at a higher amount ofcatalyst of 30 wt % (run 21) the yield increased significantlyto 88.2%. This is because H2SO4 plays the role of catalyst notonly in the reactive extraction process but also in acceleratingthe extraction of oil from jatropha seed as it was reportedthat lipid or oil dissolved better in acidic solvent (17).Therefore, by increasing the amount of H2SO4 in the reactionmixture, more oil can be extracted more quickly and easilyand be transesterified to FAME. The acidic nature of theproduct mixture due to usage of H2SO4 can be easily overcome

FIGURE 1. Individual effect of reaction time on FAME yield inreactive extraction of Jatropha seed.

FIGURE 2. Individual effect of reaction temperature on FAMEyield in reactive extraction of Jatropha seed.

FIGURE 3. Individual effect of methanol to seed ratio on FAMEyield in reactive extraction of Jatropha seed.

FIGURE 4. Individual effect of catalyst loading on FAME yieldin reactive extraction of Jatropha seed.

FIGURE 5. Interaction effect between reaction temperature andcatalyst loading on FAME yield shown as (a) a three-dimensionalplot and (b) a two-dimensional plot.

4364 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 11, 2010

in the industry by washing using alkaline water (NaOHsolution). Then the resulting neutral washed water can bereused for the subsequent washing cycle. This can signifi-cantly reduce the usage of water, making it more environ-mentally friendly.

Interaction between Parameters. As shown in Table 4,among all of the interaction terms, only the term BD issignificant at the 95% confidence level. This means that onlythe interaction between reaction temperature (B) and catalystloading (D) affects the yield of FAME significantly. The three-dimensional and two-dimensional plots for the interactionbetween reaction temperature and catalyst loading are shownin Figure 5, panels a and b, respectively. It can be seen thatwhen the catalyst loading is set at a lower value (11.25 wt %),there is only a marginal increase in the yield of FAME withhigher reaction temperature. However, when the catalystloading is set at a higher value (23.75 wt %), the yield ofFAME increases more significantly with higher reactiontemperature. This finding further strengthens our claim thatcatalyst loading is a more prominent process parameter thataffects the yield of FAME as compared to reaction temper-ature (higher F value for catalyst loading compared to reactiontemperature as shown in Table 4). This can be justified astransesterification of oil to FAME is a catalyst-dependentreaction (18), in which the transesterification reaction canproceed well even at room temperature, provided sufficientcatalyst is available (19). The importance of the presence ofcatalyst is that it can lower the activation energy of a chemicalreaction (such as the transesterification reaction) so that thereaction can proceed at lower temperature, but withoutsufficient catalyst, the transesterification reaction rate wouldbe very slow even at high temperature (18).

Optimization of Process Parameters. Up to this point,the results revealed that for the reactive extraction process,all of the process parameters studied significantly affect theyield of FAME. Thus, the next step is to optimize the processparameters to obtain the highest yield of FAME. Theoptimization procedure used must be able to take into

account not only the individual process parameter but alsothe interaction between parameters. This was carried outwith the help of the optimization function embedded inDesign-Expert software using the regression model developedas shown in eq 2. Table 5 shows the limits for each of theprocess parameters used in the optimization procedure. Thesoftware predicted (based on the 95% confidence interval)that an optimum FAME yield of 99.8 wt % can be obtainedwith the following process parameters; reaction temperatureof 60 °C, methanol to seed ratio of 10.5 mL/g, 21.8 wt % ofH2SO4, and reaction period of 10 h. The predicted optimumyield was then verified by carrying out three repeatedexperimental runs using the suggested optimum condition.The repeated experiments gave an average optimum yieldof 98.1 wt %, which is very close to the predicted value (<5%error), indicating that the predicted optimum process condi-tions are valid for this study. It is noted here that althoughthe optimum yield was slightly lower (1.44%) than the yieldobtained in run 6 (Table 3), the optimum yield can beobtained at milder reaction conditions requiring less metha-nol usage (20% reduction) and reaction time (23% reduction).Besides that, the time requirement for reactive extractionwas successfully reduced to 10 from 24 h (which was obtainedin our previous study) while maintaining a high yield of almost100% (11).

Although the reaction period required for optimum yieldof FAME is 10 h, which is rather long as compared toconventional transesterification process, it should be notedthat in reactive extraction, both extraction and transesteri-fication occur simultaneously. In a conventional FAME orbiodiesel production process, oil is initially extracted fromthe seed; for example, extraction of oil from jatropha seedrequires 8 h (20), and subsequently the oil is transesterifiedto FAME, which requires an additional 5 h using two steps,acid-base-catalyzed transesterification (21). This brings thetotal processing period to 13 h, which is still 30% more thanthe time required for reactive extraction. Therefore, reactiveextraction can actually reduce the time required for biodieselproduction from various oilseeds as compared to theconventional method, and this will eventually bring downthe cost of biodiesel production.

It was found that the optimum FAME yield obtained inthis study was higher than in other similar studies reportedin the literature (14, 22). Su et al. reported (22) the use ofreactive extraction for fatty acid ester production with J. curcasL. seed as feedstock, but methyl acetate was used as reactionsolvent and enzyme as catalyst instead of methanol andH2SO4, respectively (used in this study). The yield of FAMEreported was much lower than those reported in this study.In their study, the yield obtained using 12 h of reaction time

TABLE 5. Constraints Used To Optimize FAME Yield in ReactiveExtraction

variable goallowerlimit

upperlimit

A: reaction time (h) in range 6.75 18.25B: reaction temperature (°C) in range 30 60C: methanol to seed ratio (g/mL) in range 8.75 16.25D: catalyst loading (wt %) in range 11.25 23.75Y: FAME yield (wt %) in range 95 100

TABLE 6. Comparison of Optimum Process Conditions for Jatropha Biodiesel Production Using Various Processing Technologies

process parameterreactive

extraction

homogeneous two-stepacid-base-catalyzed

transesterification (21)lipase-catalyzed

transesterification (23)supercritical

transesterification (24)

reaction temperature (°C) 60 50 °C for acid esterification65 °C for basetransesterification

50 320 °C with pressureof 8.4 MPa

methanol to oil ratio 10.5 mL/g(methanol toseed ratio)

60% w/w for acid esterification24% w/w for basetransesterification

3 (molar ratio) 43 (molar ratio)

catalyst loading (wt %) 21.8 1% w/w H2SO4 for acidesterification

1.4% w/w NaOH for basetransesterification

100

reaction time (h) 10 3 h for acid esterification2 h for basetransesterification

24 4 min

optimum yield (wt %) 98.1 90 22.5 100

VOL. 44, NO. 11, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 4365

was only 63.9 wt %, but in this study, 98.1 wt % FAME yieldcan be achieved in only 10 h of reaction time (22). In anotherstudy, municipal primary and secondary sludges were used

as lipid feedstock for the production of biodiesel using acid-catalyzed reactive extraction. The maximum reported FAMEyields were merely 14.5 and 2.5% for primary and secondary

TABLE 7. Step-by-Step Unit Operation Comparison for Production of Biodiesel from Jatropha Using Reactive Extraction andConventional Process

4366 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 11, 2010

sludges, respectively, even at 24 h of reaction time (14). Thisreveals the fact that J. curcas L. seeds are actually a verypromising feedstock for the production of biodiesel usingacid-catalyzed reactive extraction as compared to municipalsludges because these wastes could contain a lot of impuritiesthat might affect the reactive extraction process (11).

Table 6 shows the comparison of optimum processconditions for various biodiesel processing technologies.Although homogeneous two-step acid-base-catalyzed trans-esterification requires less time and catalyst, the need for atwo-step process, higher methanol usage, and lower yieldare not favorable. On the other hand, although lipase-catalyzed transesterification requires less methanol as com-pared to reactive extraction, this enzymatic process needshigh catalyst loading as well as reaction time and the FAMEyield obtained was significantly lower. Supercritical trans-esterification can achieve a very high FAME yield in a shortreaction time, but the requirement of very high reactiontemperature and reaction pressure (50 times higher thanreactive extraction and other processing technologies) madethis technology unfavorable from an industrial point of viewas a large amount of energy is required. In addition, specialhandling, especially from the safety aspect, is also requiredfor the supercritical transesterification process.

An attempt was also made to compare the step-by-stepunit operations for producing biodiesel from jatropha usingthe conventional homogeneous two-step acid-base-cata-lyzed transesterification and reactive extraction as shown inTable 7. It can be clearly shown that reactive extractionreduced the overall biodiesel processing steps owing to theexclusion of the oil extraction step and a simpler transes-terification process. In short, acid-catalyzed reactive extrac-tion can be a breakthrough technology for biodiesel pro-duction from oilseed that contains high FFA such as J. curcasL. using a single-step method.

AcknowledgementWe acknowledge the funding given by Universiti SainsMalaysia (Research University Grant 1001/PJKIMIA/814062,Short Term Grant 304/PJKIMIA/6039015, Research UniversityPostgraduate Research Grant Scheme 1001/PJKIMIA/8031017), and the Ministry of Science, Technology andInnovation, Malaysia (National Science Fellowship), for thisproject.

Literature Cited(1) Malaysia biofuels annual report 2009; available at http://

gain.fas.usda.gov/Recent%20GAIN%20Publications/General%20Report_Kuala%20Lumpur_Malaysia_6-12-2009.pdf.

(2) Om Tapanes, N. C.; Gomes Aranda, D. A.; de Mesquita Carneiro,J. W.; Ceva Antunes, O. A. Transesterification of Jatropha curcasoil glycerides: theoretical and experimental studies of biodieselreaction. Fuel 2008, 87, 2286–2295.

(3) Nelson, R. G.; Schrock, M. D. Energetic and economic feasibilityassociated with the production, processing, and conversion ofbeef tallow to a substitute diesel fuel. Biomass Bioenerg. 2006,30, 584–591.

(4) Chhetri, A. B.; Chris Watts, K.; Rafiqul Islam, M. Waste cookingoil as an alternate feedstock for biodiesel production. Energies2008, 1, 3–18.

(5) Kansedo, J.; Lee, K. T.; Bhatia, S. Cerbera odollam (sea mango)oil as a promising non-edible feedstock for biodiesel production.Fuel 2009, 88, 1148–1150.

(6) Prueksakorn, K.; Gheewala, S. H. Full chain energy analysis ofbiodiesel from Jatropha curcas L in Thailand. Environ. Sci.Technol. 2008, 42, 3388–3393.

(7) Jatropha shines as non-food oil for biodiesel 2008; available athttp://www.reuters.com/article/GlobalAgricultureandBiofuels08/idUSHKG28600520080115.

(8) Lam, M. K.; Tan, K. T.; Lee, K. T.; Mohamed, A. R. Malaysianpalm oil: surviving the food versus fuel dispute for a sustainablefuture. Renew. Sust. Energ. Rev. 2009, 13, 1456–1464.

(9) Mission biofuel primes for biodiesel action 2008; available at http://www.energycurrent.com/index.php?id)2&storyid)10461.

(10) Zeng, J.; Wang, X.; Zhao, B.; Sun, J.; Wang, Y. Rapid in situtransesterification of sunflower oil. Ind. Eng. Chem. Res. 2009,48 (2), 850–856.

(11) Shuit, S. H.; Lee, K. T.; Kamaruddin, A. H.; Yusup, S. Reactiveextraction and in situ transesterification of Jatropha curcas L.seeds for the production of biodiesel. Fuel 2010, 89, 527–530.

(12) Haas, M. J; Scott, K. M. Moisture removal substantially improvesthe efficiency of in situ biodiesel production from soybeans.J. Am. Oil Chem. Soc. 2007, 84, 197–204.

(13) Haas, M. J.; Scott, K. M.; Foglia, T. A.; Marmer, W. N. The generalapplicability of in situ transesterification for the production offatty acid esters from a variety of feedstocks. J. Am. Oil Chem.Soc. 2007, 84, 963–970.

(14) Mondala, A.; Liang, K.; Toghiani, H.; Hernandez, R.; French, T.Biodiesel production by in situ transesterification of municipalprimary and secondary sludges. Bioresour. Technol. 2008, 100,1203–1210.

(15) Siler-Marinkovic, S.; Tomassevic, A. Transesterification ofsunflower oil in situ. Fuel 1998, 17 (12), 1389–1391.

(16) Samart, C.; Sreetongkittikul, P.; Sookman, C. Heterogeneouscatalysis of transesterification of soybean oil using Kl/meso-porous silica. Fuel Process. Technol. 2009, 90 (7-8), 922–925.

(17) Hensarling, T. nP.; Jacks, T. J. Solvent extraction of lipids fromsoybeans with acidic hexane. J. Am. Oil Chem. Soc. 1983, 80 (4),783–784.

(18) Li, H.; Xie, W. Transesterification of soybean oil to biodieselwith Zn/I2 catalyst. Catal. Lett. 2006, 107 (1-2), 25–30.

(19) Haas, M. J.; Scott, K. M.; Marmer, W. N.; Foglia, T. A. In situalkaline transesterification: an effective method for the produc-tion of fatty acid esters from vegetable oils. J. Am. Oil Chem.Soc. 2004, 81 (1), 83–89.

(20) Sayyar, S.; Abidin, Z. Z.; Yunus, R.; Muhammad, A. Extractionof oil from Jatropha seeds-optimization and kinetics. Am. J.Appl. Sci. 2009, 6 (7), 1390–1395.

(21) Berchmans, H. J.; Hirata, S. Biodiesel production from crudeJatropha curcas L. seed oil with a high content of free fattyacids. Bioresour. Technol. 2008, 99, 1716–1721.

(22) Su, E.; Xu, W.; Gao, K.; Zheng, Y.; Wei, D. Lipase-catalyzed insitu reactive extraction of oilseeds with short-chained alkylacetates for fatty acid ester production. J. Mol. Catal. B: Enzym.2007, 48, 28–32.

(23) Rattanaphra, D.; Srinophakun, P. Biodiesel production fromcrude sunflower oil and crude Jatropha oil using immobilizedlipase. J. Chem. Eng. Jpn. 2010, 43 (1), 104–108.

(24) Hawash, S.; Kamal, N.; Zaher, F.; Kenawi, O; Diwani, G. E.Biodiesel fuel from Jatropha oil via non-catalytic supercriticalmethanol transesterification. Fuel 2009, 88, 579–582.

ES902608V

VOL. 44, NO. 11, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 4367