Fuel Processing Technology 91 (2010) 1274–1281

Contents lists available at ScienceDirect

Fuel Processing Technology

j ourna l homepage: www.e lsev ie r.com/ locate / fuproc

Continuous production of soybean biodiesel with compressed ethanol in amicrotube reactor

Camila da Silva a,b,⁎, Fernanda de Castilhos c,d, J. Vladimir Oliveira d, Lucio Cardozo Filho b

a Department of Technology, Maringá State University, (UEM), CEP: 87506-370, Umuarama, Brazilb Department of Chemical Engineering, Maringa State University, (UEM). Av. Colombo 5790, Maringa, PR, 87020-900, Brazilc Department of Chemical Engineering, Paraná Federal University (UFPR), Polytechnic Center (DTQ/ST/UFPR), Jardim das Américas, Curitiba, PR, 82530-990, Brazild Department of Food Engineering, URI, Campus de Erechim, Av. Sete de Setembro, 1621, Erechim, RS, 99700-000, Brazil

⁎ Corresponding author. Department of Technolog(UEM). Av. Angelo Moreira da Fonseca, Umuarama, PR44 36219339; fax: +55 44 36219300.

E-mail address: camiladasilva.eq@gmail.com (C. da S

0378-3820/$ – see front matter © 2010 Elsevier B.V. Adoi:10.1016/j.fuproc.2010.04.009

a b s t r a c t

a r t i c l e i n f o

Article history:Received 24 August 2009Received in revised form 11 March 2010Accepted 19 April 2010

Keywords:Supercritical biodieselEthyl estersContinuous processCatalyst-free process

This work investigates the production of fatty acid ethyl esters (FAEEs) from the transesterification ofsoybean oil in supercritical ethanol in a continuous catalyst-free process. Experiments were performed in amicrotube reactor in the temperature range of 523 K to 598 K, from 10 MPa to 20 MPa, varying the oil toethanol molar ratio from 1:10 to 1:40, and evaluating the effects of addition of carbon dioxide as co-solvent.Results showed that ethyl esters yield obtained in the microtube reactor (inner diameter 0.76 mm) werehigher than those obtained in a tubular reactor (inner diameter 3.2 mm) possibly due to improved mass-transfer conditions attained inside the microtube reactor. Non-negligible reaction yields (70 wt.%) wereachieved along with low total decomposition of fatty acids (b5.0 wt.%). It is shown that the use of carbondioxide as co-solvent in the proposed microtube reactor did not significantly affect the ethyl esters yieldwithin the experimental variable ranges investigated.

y, Maringa State University,, 87506-370, Brazil. Tel.: +55

ilva).

ll rights reserved.

© 2010 Elsevier B.V. All rights reserved.

1. Introduction

The merits of biodiesel (fatty acid ethyl or methyl esters) obtainedfrom vegetable oils or animal fats as an alternative to mineral dieselcomprise a non-toxic, biodegradable, domestically produced andrenewable source is well documented in the literature [1–4]. Becauseof the well-known environmental and economical benefits, biodieselfuel may be expected as a good alternative to petroleum-based fuel.

Transesterification, among other processes used for biodieselproduction, has been the most common way to produce biodiesel [1,2].Conventionally, transesterification can be performed using alkaline, acid,enzyme catalysts or heterogeneous chemical catalysts [1,2,5–7].

Recently, a free-catalyst technique for the transesterification ofvegetable oils using an alcohol at supercritical conditions has beenproposed as an alternative method for biodiesel production [8–16].According to the current literature, catalyst-free alcoholysis reactionsat high temperature and pressure conditions provide improved phasesolubility, decrease mass-transfer limitations, afford higher reactionrates and make easier separation and purification steps of theproducts. Besides, it has been shown that the so-called supercriticalmethod is more tolerant to the presence of water and free fatty acidsthan the conventional alkali-catalyzed technique, and hence more

tolerant to various types of vegetable oils, even for fried andwaste oils[17–19]. Thus, the supercritical method appears to be attractive forapplication in continuous mode, which is the primary importance toassure a competitive cost to biodiesel fuel [20–24].

However, supercritical method requires high molar ratio of alcoholto oil and high temperature and pressure conditions for the reaction topresent satisfactory yield levels, leading to high processing costs andcausing in many cases the degradation of the fatty acids esters formed[25], hence decreasing the reaction yield [20–22,24]. Attempts toreduce the expected high operating cost and product degradation havebeen made through the addition of co-solvents [26–30], two-stepprocess with removal of glycerol generated in the first step [31] andadopting a two-step process comprising hydrolysis of triglycerides insubcritical water and subsequent esterification of fatty acids [20,21].

Another approach suitable for the biodiesel production insupercritical conditions is the use of microreactor systems operatingin continuous mode. Microreactor systems designed for continuousproduction have been studied in recent years for the transesterifica-tion of vegetable oils [32–36]. In the microreactor system, mass andheat transfer could be greatly intensified due to its small space with alarge surface area-to-volume ratio [34], providing high process yieldsin low reaction times [33], that can in supercritical transesterificationprevent the decomposition of fatty acids and decrease the operatingconditions. However, studies available in the open literature refer tobase-catalyzed alcoholysis reactions conducted at low pressure, withno reports found on the use of microreactors for the transesterifica-tion reaction at sub- and supercritical alcohol conditions.

1275C. da Silva et al. / Fuel Processing Technology 91 (2010) 1274–1281

In this sense, the main objective of the present work is toinvestigate the continuous transesterification of soybean oil undersupercritical ethanol conditions in a microtube reactor. For thispurpose it was assessed the effects of temperature in the range of523 K to 598 K, pressure from 10 MPa to 20 MPa, oil to ethanol molarratio from 1:10 to 1:40 and the addition of carbon dioxide as co-solvent for CO2 to ethanol molar ratio ranging from 1:5 to 1:10.

2. Materials and methods

2.1. Materials

Commercial refined soybean oil (Soya) and ethanol (Merck 99.9%)were used as substrates without further treatment and carbon dioxideas co-solvent (White Martins, 99%). Other solvents, standards andreagents used in the derivatization step required for the analysis weresupplied by Sigma-Aldrich. Chemical composition for the soybean oilused in thiswork is reported elsewhere [37]. The acid value (mgKOH/g)and water content (wt.%, Karl Fischer titration method, DL 50, Mettler-Toledo) were determined to be approximately 0.2 and 0.04,respectively.

2.2. Apparatus and experimental procedure

The experimental reaction system used in this work, schematicallypresented in Fig. 1, is similar to that used previously by Silva et al. [24],with the exception of the modification for carbon dioxide addition.Transesterification reactions were carried out in duplicate using amicrotube reactor with a capacity of 36.5 mL made of stainless steeltubing (316 L 1/16in. OD internal diameter 0.76 mm HIP). Thesubstrates, ethanol and oil, were placed in a closed erlenmeyer andmixed by means of a mechanical stirring device and then were fedinto the reaction system by a high-pressure liquid pump (Acuflow).Co-solvent (carbon dioxide) was added to the system at a pre-established flow rate using a syringe pump (Isco, model 500D). Themicrotube reactor was placed in a furnace with controlled temper-ature and monitored by two thermocouples directly connected at the

Fig. 1. Schematic diagram of the experimental apparatus. RM— reactional mixture; MS—mereservoir; B — thermostatic baths; SP — syringe pump; F — furnace; TR — tubular reactor; T1outlet; DA— data acquisition system; CS— cooling system; V1— feed valve; PI— pressure ind

inlet and outlet of the reactor. With this arrangement, the reactiontemperature was controlled with a precision better than 5 K.

The system pressure was controlled by a control loop composed bya pressure transducer (Smar, model A5), a PID controller (Novus,Model N1100) and an electropneumatic valve (Baumann™, model51000). Though more complex definitions can be used for theevaluation of the time spent in the reactor (the residence time)[21,24], in this work the residence time was computed dividing thevolume of the reactor (mL) by the flow rate of substrates (mL/min) setin the liquid pump.

Samples were collected periodically in a glass vial placed at thereactor outlet after reaching the steady state condition, i.e., after areactor space–time had been elapsed at least three times. Preliminarytests were carried out for some experimental conditions, affordingexcellent reproducibility of the experimental apparatus [38]. Based onduplicate experiments, the overall experimental error was found to beless than 5% on FAEE yield.

2.3. Gas chromatography (GC) analysis of fatty acid ethyl esters (FAEEs)

Samples were first submitted to ethanol evaporation to constantweight in a vacuum oven (338 K, 0.05 MPa) and then diluted with2 mL of ethanol and 8 mL of n-heptane. Afterwards, a little amountwas transferred to a 1 mL flask in order to obtain a concentration of1000 ppm and then it was added the internal standard at aconcentration of 250 ppm using n-heptane as solvent. After that,1 µL of solution was injected in triplicate in the gas chromatograph(Shimadzu GC-2010), equipped with FID, autoinjector AOC-20i and acapillary column (DBWAX, 30 m×0.25 mm×0.25 µm). Column tem-perature was programmed from 393 K, holding 2 min, heating to453 K at 10 K/min, holding 3 min, and to 503 K at 5 K/min, holding2 min. Helium was used as carrier gas, and the injection and detectortemperatures were 523 K with split ratio: 1:50.

Compounds were quantified upon analysis following the standardUNE-EN 14103 [39] and FAEE yield was then calculated based on thecontent of ethyl esters in the analyzed sample and on the reactionstoichiometry.

chanical stirring device; LP— high-pressure liquid pump; CV— check-valve; A— solvent— temperature indicator at the reactor inlet; T2 — temperature indicator at the reactoricator; PIC— controller; V2— pressure control valve; S— glass collector; G— gas output.

1276 C. da Silva et al. / Fuel Processing Technology 91 (2010) 1274–1281

2.4. Decomposition of fatty acid

Sampleswere treatedwith BF3/MeOH [40] in order to derivatize allthe fatty acids (mono, di- and triglycerides, free fatty acids and alsoethyl esters) to the corresponding methyl esters, and then analyzedby GC as described above. For the evaluation of the degradationpercentage, palmitic acid was assumed not liable to degradation,considering its high stability [22,25]. Degradation was thus calculatedaccording to the following equation, as described by Vieitez et al.[37,41]:

decomposition %ð Þ = 100 × 1− ∑PiP16:0

� �s×

P16:0∑Pi

� �o

� �ð1Þ

where ΣPi is the summation of all fatty acid methyl ester percentages,P16:0 is the percentage of 16:0 ethyl ester, and subscripts “s” and “o”indicate that the expressions between brackets were evaluatedconsidering the composition of the sample product and the originaloil, respectively.

2.5. Analysis of mono-, di- and triglycerides

Compounds were quantified upon analysis following the stan-dard UNE-EN 14105 [42]. Samples prepared as described above weretreated with MSTFA/pyridine to ensure derivatization of free al-cohols to their corresponding trimethylsilyl esters and transferred toa 10 mL flask using n-heptane as solvent. GC analysis was con-ducted in the above-described equipment, with a column HT5(25 m×0.32 mm×0.1 µm) and on-column injector. Helium was thecarrier gas, and oven temperature was programmed following thesteps recommended in the standard.

3. Results and discussion

3.1. Comparison of microtube with tubular reactor

To evaluate the influence of the microtube reactor (inner diameterof 0.76 mm) on the FAEE yield, experiments were performed keepingthe pressure fixed at 20 MPa, the molar ratio of oil to ethanol at 1:20and residence time at 25 min, varying the temperature in the rangefrom 523 to 598 K. Fig. 2 shows the effects of inner diameter on theFAEE yield, comparing the results obtained in this work with thosereported by Silva et al. [24] for the same residence time using a tubularreactor (inner diameter of 3.2 mm). It can be seen from this figure thatat the lowest temperature (523 K) only 3.12 wt.% FAEE yield is

Fig. 2. Effect of inner reactor diameter and temperature on the FAEE yield at 20 MPa,25 min and oil to ethanol molar ratio of 1:20.

obtained in the tubular reactor, while 19 wt.% is reached using themicrotube reactor. At 598 K this yield is increased from 38 wt.% to53 wt.% when passing from the tubular to themicrotube reactor at thesame residence time. Such results demonstrate that higher ethylesters yields can be achieved at lower temperatures, small reactiontimes with a smaller reactor inner diameter, and minimize the totaldecomposition of fatty acids.

In the current literature no reports were found on the use of microreactors or microtube reactors to conduct the supercritical transes-terification of vegetable oils. Sun et al. [33] showed that the use ofreactors with inner diameters of 0.25 and 0.53 mm provided higheryields at lower times for alkaline alcoholysis of sunflower oil inmethanol at low pressure conducted in capillary microreactors, as aconsequence of the larger specific surface area of the capillary withsmaller dimensions and intensified mass transfer. Guan et al. [34,35]demonstrated the strong influence of the reactor diameter on theesters production through alkaline transesterification of vegetableoils, with high yields achieved in microtubes of 0.4 mm compared tothe results from a reactor with internal diameter of 1.0 mm.

3.2. Effect of temperature

The effect of temperature on the ethyl esters yield was assessedkeeping the oil to ethanol molar ratio fixed at 1:20, pressure at20 MPa, varying the temperature from 523 K to 598 K. One canobserve from Fig. 3 that an increase in temperature led to a sharpenhancement of FAEE yield and faster initial reaction rates. At 523 Kyields in the order of 19 wt.% was obtained for 25 min of reaction,while 53 wt.% was obtained for the same period for supercriticaltreatment at 598 K. The ethyl esters yield increased with residencetime for all conditions studied, with no observed decrease in reactionyield at larger reaction times, as observed in the works of Minami andSaka [21], He et al. [22] and Silva et al. [24]. For the temperatures of573 and 598 K the experimental decomposition values of 1.1 and4.8 wt.% were determined at 45 min, respectively. Therefore, FAEEyields in the order of 70 wt.% to ethyl esters could be obtained at 598 Kwith low decomposition of the constituents of the reaction medium.

Gui et al. [43] reported ester yields of 79.2 wt.% for the batch-modetransesterification of palm oil in supercritical ethanol at 623 K usingand oil to ethanol molar ratio of 1:33 and Tan et al. [44] obtained70 wt.% of yield in methyl esters at 633 K and oil to methanol molarratio of 1:30 using a batch-type tube reactor. Minami and Saka [21]reported 80 wt.% of esters for a continuous process using methanol at623 K, 20 MPa and oil:methanol molar ratio of 1:42. He et al. [22]

Fig. 3. Effect of temperature on the FAEE yield at 20 MPa using oil to ethanol molar ratioof 1:20. Symbols are experimental data and continuous lines are provided just toimprove visualization.

1277C. da Silva et al. / Fuel Processing Technology 91 (2010) 1274–1281

achieved yields in the order of 77 wt.% in methyl esters in thealcoholysis of soybean oil at supercritical conditions of 583 K, 35 MPaand oil:methanol molar ratio of 1:40. For the continuous transester-ification of soybean oil in supercritical ethanol, Silva et al. [24]reported approximately 80 wt.% of reaction conversion to ethyl estersat 623 K, 20 MPa and oil to ethanol molar ratio of 1:40; at in thiscondition Vieitez et al. [37] reported nearly 18 wt.% total decompo-sition of fatty acids.

Fig. 4 shows the results in terms of percentage content oftriglycerides (TGs), diglycerides (DGs), monoglycerides (MGs) andglycerol, for the temperatures of 573 and 598 K, at different residencetimes. Comparing the results presented in Fig. 4a and b it seems thattemperature directly influences the reaction rate of other componentsof the reaction medium of supercritical alcoholysis. Consumption ofTG was favored with increasing temperature, in approximately15 min the percentage of TG is 45 wt.% at 573 K and 25 wt.%at 598 K, and an increase in residence time (45 min) leads to only7.6 wt.% and 4.7 wt.% for both temperatures studied, demonstratingthe high conversion of TG under these conditions. As one shouldexpect, high contents of DG and MG are observed in the beginning ofthe reaction, showing the evident formation and consumption of bothsubstances as reaction takes place. At 598 K for longer residencetimes, the whole percentage did not exceed 12 wt.% for thesecomponents. One should also note from Fig. 4b that, initially, the

Fig. 4. Content of triglycerides, diglycerides, monoglycerides and glycerol in theproducts obtained at 20 MPa, oil to ethanol molar ratio of 1:20 and temperature of:(a) 573 K; (b) 598 K.

content of the intermediate reaction product, DG, is higher than MGbut, consistently, as reaction involves the content of DG is progres-sively surpassed by the final reaction product.

The glycerol content increases with residence time, with nodecomposition observed for this compound, as reported in the worksof Anistescu et al. [30] and Aimaretti et al. [45]. In 45 min reaction aglycerol content of 4.9 wt.% at 573 K and 6.6 wt.% at 598 K wasobserved, corresponding to FAEE yield of 52 wt.% and 70 wt.%,respectively, shown in Fig. 3. The highest yield in terms of ethylesters (approx. 70 wt.%), shown in Fig. 4b, was obtained at 598 K,corresponding to a TG content of only 4.7 wt.% and intermediateproducts (DG and MG) of 11.2 wt.%.

The water content and percentage of free fatty acid during thecourse of the ethanolysis reaction was not determined, but we believethat due to the low water and free fatty acid (FFA) contents initiallypresent in the oil, these parameters did not have a relevant effect onthe fatty acid ethyl esters (FAEEs) yield obtained, as also verified insome studies reported in the literature [17,37,41].

3.3. Effect of pressure

The effect of pressure on the alcoholysis reaction was evaluatedadopting the oil to ethanol molar ratio of 1:20 and temperatures of573 K and 598 K. The pressure values of 10 MPa, 15 MPa and 20 MPawith results shown in Fig. 5a and b was considered. The pressurerange investigated in this work was based on previous studies

Fig. 5. Effect of pressure on the FAEE yield for oil to ethanol molar ratio of 1:20 at:(a) 573 K; (b) 598 K.

Fig. 6. Content of triglycerides, diglycerides, monoglycerides and glycerol in theproducts obtained at 573 K, oil to ethanol molar ratio of 1:20 and pressure of:(a) 10 MPa; (b) 15 MPa.

1278 C. da Silva et al. / Fuel Processing Technology 91 (2010) 1274–1281

conducted by our working group [24,37,41] and on reports availablein the literature [8–11,21,22]. Pressures above 20 MPa were notconsidered in this work due to the low increase in FAEE yields and thehigh initial investments (equipment costs) for the implementation ofsuch process operated at higher pressures. It is believed that operationpressures above 20 MPa may not be industrially viable, increasing theimplementation costs for biodiesel production through supercriticaltransesterification method.

It can be seen from Fig. 5 the best FAEE yields were obtained for20 MPa, consistent with the results presented in the literature in theworks of Kusdiana and Saka [8–10] and Demirbas [11] for transester-ification reactions in batch mode and Minami and Saka [21] and Silvaet al. [24] in continuous mode. At 573 K and 45 min of residence time,52 wt.% of ethyl esters is obtained at 20 MPa and 29 wt.% at 10 MPa. At598 K on the other hand, for the same residence time, 70 wt.% of yieldis obtained at 20 MPa and 58 wt.% at 15 MPa. These results should beconsidered together with initial investments (equipment costs) andenergy balance for possible implementation of a cost-effectiveimplementation of a technological attractive supercritical transester-ification process.

Just a few works in the literature investigate the pressure effect onthe transesterification yield using the continuous supercriticalmethod. For example, He et al. [22] evaluated the effect of pressureon the continuous transesterification of soybean oil with methanoland observed a positive effect in the range of 10 MPa to 40 MPa, withthe best condition found at 35 MPa. The authors report esters yield inthe order of 55 wt.% and 60 wt.% for the reactions conducted at20 MPa, respectively for temperatures of 573 K and 593 K in oil tomethanol molar of 1:40 at 25 min. Conversely, Bunyakiat et al. [46]observed that pressure did not affect the transesterification conver-sion of coconut oil with supercritical methanol, reporting a yield of65.82 wt.% to methyl esters for oil:methanol molar ratio of 1:24 andtemperature of 623 K, at 10 MPa and 67.84 wt.% at 19 MPa.

Fig. 6a and b presents the kinetics of production of MG, DG TG andglycerol at 573 K and pressures of 10 MPa and 15 MPa, results that canbe compared to those shown in Fig. 4a for pressure of 20 MPa. Basedon the results presented in these figures it becomes even moreevident the effect of pressure on the yield of the supercritical reactionof vegetable oils. The reaction rate of TG consumption is slower atlower conditions of pressure, directly influencing the formation ofintermediate products (DG and MG), glycerol and ethyl esters; e.g., at10 MPa and 15 MPa at longer residence times (45 min) higher levelsof DG (18.6 wt.% and 15.1 wt.%) are observed, for low percentage ofglycerol formed (1.2 wt.% and 2.7 wt.%), consistent with the low yieldvalues of esters shown in Fig. 6a. In 45 min of reaction percentages of36.5 wt.%, 20.8 wt.% and 7.6 wt.% are observed for TAG at 10, 15 and20 MPa, respectively.

Fig. 7 shows the results regarding total decomposition of fattyacids determined by derivatizationwith BF3/MetOH for the conditionspresented in Fig. 6. It was observed that system pressure affects thedecomposition for both temperatures considered. Indeed, at thehigher residence time at 573 K, the decomposition of 0.5 wt.% at10 MPa increases to 1.04 wt.% at 20 MPa, whereas the correspondingvalues at 598 K are 3.9 wt.% at 10 MPa and 4.8 wt.% at 20 MPa.However, the decomposition degree obtained for both conditionsstudied was low (b5.0 wt.%) and they may be used to conduct thereaction.

3.4. Effect of soybean oil to ethanol molar ratio

To evaluate the effect of oil to ethanol molar ratio in the range of1:10 to 1:40, reactions were conducted at two conditions: 573 K and20 MPa and 598 and 15 MPa. Fig. 8 shows the time course of FAEEyield for these conditions. At 573 K and 20 MPa, Fig. 8a, the maximumyield achieved to esters was 58 wt.% at 45 min and oil to ethanolmolarratio of 1:40, while for the same residence time 52 wt.% was obtained

for oil to ethanol molar ratio of 1:20. It can be noted that a rise in themolar ratio of alcohol to oil affords better yields in shorter residencetimes, i.e., oil to ethanol molar ratio of 1:10, 41 wt.% of FAEE yield isobtained in 45 min of reaction, whereas for oil to ethanol molar ratioof 1:20, almost the same yield is obtained in 24 min. Minami and Saka[21] obtained in nearly 40 min of reaction 48 wt.% in methyl esters at573 K, 20 MPa and oil:methanol molar ratio of 1:40 for a continuousreaction of canola oil in supercritical methanol.

The molar ratio of oil to alcohol is one of the most importantvariables affecting the yield of fatty acids esters in supercriticaltransesterification because in catalyst-free reactions an increase in thealcohol to oil molar ratio should allow greater contact betweensubstrates, thus favoring reaction conversion [9]. Kusdiana and Saka[9] and Demirbas [11] evaluated the influence of molar ratio of oil tomethanol in non-catalytic transesterification in batch mode andobtained the best results in terms of conversion to esters for thecondition of oil to methanol molar ratio of 1:42. The effect of oil toethanolmolar ratio was also studied by Varma andMadras [47] for thetransesterification of castor oil in batch mode using supercriticalethanol in the temperature range of 523 to 623 K and pressure of20 MPa, where the authors reported an increase in the reactionconversion with increasing molar ratio of ethanol to oil in the range of1:10 to 1:40. He et al. [22] evaluated the effect of alcohol to oil molarratio on the continuous transesterification of soybean oil in

Fig. 7. Effect of pressure on the reaction decomposition for oil to ethanol molar ratio of1:20 at: (a) 573 K; (b) 598 K.

Fig. 8. Effect of oil to ethanol molar ratio on the FAEE yield at: (a) 573 K and 20 MPa; (b)598 K and 15 MPa.

1279C. da Silva et al. / Fuel Processing Technology 91 (2010) 1274–1281

supercritical methanol and found that at 573 K and 32 MPa, the oil tomethanol molar ratio showed a positive effect on reaction conversionin the range of 1:6–1:40. At oil to methanol molar ratio of 1:20 and1:40 yields in the order of 64 wt.% and 70 wt.% in methyl esters arereported by the authors.

It can be seen from Fig. 8b that at 598 K and 15 MPa, 58 wt.% FAEEyield is obtained for oil to ethanol molar ratio of 1:20 and 68 wt.% foroil to ethanol molar ratio of 1:40, with lower yields obtained for thistemperature at oil:ethanol molar ratio of 1:10. Wang et al. [48]obtained in the alcoholysis of soybean oil with methanol at oil:methanol molar ratio of 1:20, 15 MPa and 623 K, yields in the order of60 wt.%, while for oil to methanol molar ratio of 1:40 the authorsreport 80 wt.% conversion of soybean oil.

Verifying the effect of oil to ethanol molar ratio on the conditionsstudied (Fig. 8a and b) one notes that the molar ratio of oil:ethanol of1:40 afforded the best results in terms of yield to esters. However,satisfactory yields can be obtained at oil to ethanol molar ratio of 1:20,hence diminishing the amount of alcohol used in the process, since amolar ratio of 1:40 corresponds to a mass ratio oil:ethanol of 1:2,twice that of 1:20. Thus, for the system investigated, within thevariable ranges investigated, the oil to ethanol molar ratio of 1:20seems to be the most suitable condition for the continuous non-catalytic production of biodiesel.

Fig. 9 shows the kinetics of intermediate products for the oil toethanol molar ratio of 1:10 and 1:40, at 573 K and 20 MPa, whileresults for the molar ratio of 1:20 are shown in Fig. 4a. It can be seen

from Figs. 9 and 4a that increasing the amount of alcohol in thereaction medium increases the reaction rate of components, enhanc-ing the percentage of glycerol obtained. The consumption of TGsincreases with time andwith increasing the molar ratio of oil:ethanol:in 45 min of reaction a TG percentage of 19.8 wt.%, 7.6 wt.% and5.94 wt.% is observed, respectively, for the conditions studied, 1:10,1:20 and 1:40. The formation and consumption of DG andMG follow alogical sequence, since the DG are formed from the reaction of TG andconsumed for MG formation, and they react to produce glycerol,leading to the final reaction product, formation of esters.

Results concerning the decomposition of fatty acids for theconditions studied in Fig. 8 are shown in Fig. 10. One should noticethat the oil to ethanol molar ratio has a significant effect on thedecomposition for both conditions considered. For instance, at 573 Kthe percentage of decomposition did not exceed 1.0 wt.% for the rangeof oil:ethanolmolar ratio investigated. It can be seen from Fig. 10b thatat the temperature of 598 K, higher decomposition rates are observedfor the oil:ethanol molar ratio of 1:10, and thus at higher residencetimes (45 min) the percentage of decomposition of 5.0 wt.%, 4.8 wt.%and 4.2 wt.% are observed for oil to ethanol molar ratio of 1:10, 1:20and 1:40 respectively.

3.5. Effect of co-solvent addition

The effect of addition of carbon dioxide as co-solvent on the ethylesters yield was assessed keeping temperature at 573 K and pressure

Fig. 9. Content of triglycerides, diglycerides, monoglycerides and glycerol in theproducts obtained at 573 K, 20 MPa and oil to ethanol molar ratio of: (a) 1:10; (b) 1:40.

Fig. 10. Effect of oil to ethanol molar ratio on the reaction decomposition at: (a) 573 Kand 20 MPa; (b) 598 K and 15 MPa.

Table 1Effect of addition of carbon dioxide on the FAEE yield (wt.%) at 20 MPa and 573 K.

Run Oil toethanolmolarratio

CO2 toethanolmolarratio

Residence time (min)

20 35

1 1:10 0 18.9±0.2 35.5±0.72 1:10 1:10 19.0±0.5 36.6±0.73 1:10 1:5 18.1±0.4 35.3±0.94 1:20 0 29.1±0.1 48.2±0.35 1:20 1:10 30.6±0.5 49.7±0.56 1:20 1:5 29.7±0.9 50.0±0.87 1:40 0 39.4±0.7 54.0±1.38 1:40 1:10 40.2±0.7 54.3±0.79 1:40 1:5 39.1±0.8 54.9±0.3

1280 C. da Silva et al. / Fuel Processing Technology 91 (2010) 1274–1281

at 20 MPa, for different oil to ethanol molar ratios (1:10, 1:20 and1:40) and CO2 to ethanol molar ratios of 1:5 and 1:10, from a 23

factorial design for 20 and 35 min of residence time (Table 1). Thestatistical analysis performed using the Statistica® 6.0Softwareshowed that oil to ethanol molar ratio has a significant effect on theFAEE yield, as previously discussed. However, neither the mainvariable CO2 to ethanol molar ratio nor the cross interaction betweenthe variables studied showed a significant effect on process yield at95% confidence level.

Han et al. [27] investigated the production of biodiesel from soybeanoil using methanol and supercritical CO2 as co-solvent in batch modewith agitation, andnoticed at573 Kandoil to ethanolmolar ratioof 1:33that reaction conversion to esterswas about 75 wt.%without addition ofco-solvent, with an increase of approximately 82 wt.% for addition of0.04 (molar ratio CO2/methanol) and greater than 95 wt.% for additionof 0.08, 0.1 and 0.2 (molar ratio CO2/methanol). Authors reported alsothat conversions at conditions above 583 Kwere little influenced by theaddition of co-solvent.

Imahara et al. [49] showed that a decrease in the reactionconversion occurred when the CO2/methanol molar ratio wasincreased above 0.1 for the transesterification reaction of canola oilin a batch-mode system without agitation.

One of the few works available in the literature regarding the useof carbon dioxide in a continuous transesterification process is due tothe investigation performed by Anistescu et al. [30], which reportsexperimental data on supercritical transesterification of soybean oil in

ethanol and methanol. Nevertheless, the range of the variablesstudied in that work is different from the present contribution, andfurther results are presented solely in terms of TG conversion and notfor the reaction products (esters, glycerol, MG and DG). The work ofBertoldi et al. [50] for continuous production of soybean biodiesel insupercritical ethanol and dioxide carbon as co-solvent showed thatFAEE yield decreasedwith increasing addition of carbon dioxide to thesystem for reaction conduzed in a tubular reactor for CO2 tosubstrates mass ratio greater than 0.05:1.

1281C. da Silva et al. / Fuel Processing Technology 91 (2010) 1274–1281

4. Conclusions

This work reported experimental data on ethyl esters productionfrom soybean oil in a continuous microtube reactor, evaluating theinfluence of temperature, pressure, oil to ethanol molar ratio,residence time and addition of carbon dioxide as co-solvent. Resultsshowed that ethyl esters yields obtained in the microtube reactor(inner diameter 0.76 mm) were higher than those obtained in atubular reactor (inner diameter 3.2 mm) probably as a consequence ofthe larger specific surface area of the microtube with smallerdimensions and intensified mass transfer. In the experimental rangeinvestigated, it was verified that temperature, pressure and oil toethanol molar ratio had a positive effect on FAEE yield, andappreciable yields were obtained for different experimental condi-tions: at 598 K, 20 MPa and oil to ethanol molar ratio of 1:20 and598 K, 15 MPa and oil to ethanol molar ratio of 1:40, with low totaldecomposition of fatty acids (b5.0 wt.%) observed. The addition ofcarbon dioxide as co-solvent showed no significant effect with 95% ofconfidence on the esters yield, for the conditions studied in this work.In an attempt to improve process efficiency, further work underdevelopment by our working group comprises the use of a series ofreactors and recycle the output reactor stream.

Acknowledgements

The authors thank CNPq, PROCAD/Pro-Engenharia — CAPES,Petrobras S.A., Intecnial S.A. and URI/Campus de Erechim for thefinancial support. One of the authors (Camila da Silva) thanks CNPq(process 140933/2008-5) for the scholarship.

References

[1] H. Fukuda, A. Kondo, H. Noda, Biodiesel fuel production by transesterification ofoils, J. Biosci. Bioeng. 92 (2001) 405–416.

[2] F. Ma, M.A. Hanna, Biodiesel production: a review, Bioresour. Technol. 70 (1999)1–19.

[3] A.E. Srivastava, R. Prasad, Triglycerides-based diesel fuels, Renew. Sustain. EnergyRev. 4 (2000) 111–133.

[4] R. Altin, S. Çetinkaya, H.S. Yucesu, The potential of using vegetable oil fuels as fuelfor diesel engines, Energy Convers. Manag. 42 (2001) 529–538.

[5] L.C. Meher, D. Vidya Sagar, S.N. Naik, Technical aspects of biodiesel production bytransesterification — a review, Renew. Sustain. Energy Rev. 10 (2006) 248–268.

[6] J.M. Marchetti, V.U. Miguel, A.F. Errazu, Possible methods for biodiesel production,Renew. Sustain. Energy Rev. 11 (2007) 1300–1311.

[7] J.V. Gerpen, Biodiesel processing and production, Fuel Process. Technol. 86 (2005)1097–1107.

[8] D. Kusdiana, S. Saka, Biodiesel fuel from rapeseed oil as prepared in supercriticalmethanol, Fuel 80 (2001) 225–231.

[9] D. Kusdiana, S. Saka, Kinetics of transesterification in rapeseed oil to biodiesel fuelas treated in supercritical methanol, Fuel 80 (2001) 693–698.

[10] D. Kusdiana, S. Saka, Methyl esterification of free fatty acids of rapeseed oil astreated in supercritical methanol, J. Chem. Eng. Japan 34 (2001) 383–387.

[11] A. Demirbas, Biodiesel from vegetable oils via transesterification in supercriticalmethanol, Energy Convers. Manag. 43 (2002) 2349–2356.

[12] G. Madras, C. Kolluru, R. Kumar, Synthesis of biodiesel in supercritical fluids, Fuel83 (2004) 2029–2033.

[13] Y. Warabi, D. Kusdiana, S. Saka, Reactivity of triglycerides and fatty acids ofrapeseed oil in supercritical alcohols, Bioresour. Technol. 91 (2004) 283–287.

[14] A. Demirbas, Biodiesel production via non-catalytic SCF method and biodiesel fuelcharacteristics, Energy Convers. Manag. 47 (2006) 2271–2282.

[15] T. Pinnarat, P. Savage, Assessment of noncatalytic biodiesel synthesis usingsupercritical reaction conditions, Ind. Eng. Chem. Res. 47 (2008) 6801–6808.

[16] D. Wen, H. Jiang, K. Zhang, Supercritical fluids technology for clean biofuelproduction, Prog. Nat. Sci. 19 (2009) 273–284.

[17] D. Kusdiana, S. Saka, Effects of water on biodiesel fuel production by supercriticalmethanol treatment, Bioresour. Technol. 91 (2004) 289–295.

[18] V. Rathore, G. Madras, Synthesis of biodiesel from edible and non-edible oils insupercritical alcohols and enzymatic synthesis in supercritical carbon dioxide,Fuel 86 (2007) 2650–2659.

[19] A. Demirbas, Biodiesel from waste cooking oil via base-catalytic and supercriticalmethanol transesterification, Energy Convers. Manag. 50 (2009) 923–927.

[20] D. Kusdiana, S. Saka, Two-step preparation for catalyst-free biodiesel fuelproduction, Appl. Biochem. Biotechonol. 113 (2004) 781–791.

[21] E. Minami, S. Saka, Kinetics of hydrolysis and methyl esterification for biodieselproduction in two-step supercritical methanol process, Fuel 85 (2006)2479–2483.

[22] H. He, W. Tao, S. Zhu, Continuous production of biodiesel from vegetable oil usingsupercritical methanol process, Fuel 86 (2007) 442–447.

[23] J.M.N. van Kasteren, A.P. Nisworo, A process model to estimate the cost ofindustrial scale biodiesel production from waste cooking oil by supercriticaltransesterification, Resour. Conserv. Recycl. 50 (2007) 442–458.

[24] C. Silva, T.A. Weschenfelder, S. Rovani, F.C. Corazza, M.L. Corazza, C. Dariva, J.V.Oliveira, Continuous production of fatty acid ethyl esters from soybean oil incompressed ethanol, Ind. Eng. Chem. Res. 46 (2007) 5304–5309.

[25] H. Imahara, E. Minami, S. Hari, S. Saka, Thermal stability of biodiesel insupercritical methanol, Fuel 87 (2007) 1–6.

[26] W. Cao, H. Han, J. Zhang, Preparation of biodiesel from soybean oil usingsupercritical methanol and co-solvent, Fuel 84 (2005) 347–351.

[27] H. Han, W. Cao, J. Zhang, Preparation of biodiesel from soybean oil usingsupercritical methanol and CO2 as co-solvent, Process. Biochem. 40 (2005)3148–3151.

[28] P. Hegel, G. Mabe, S. Pereda, E.A. Brignole, Phase transitions in a biodiesel reactorusing supercritical methanol, Ind. Eng. Chem. Res. 46 (2007) 6360–6365.

[29] J.Z. Yin, M. Xiao, J.B. Song, Biodiesel from soybean oil in supercritical methanolwith co-solvent, Energy Convers. Manag. 49 (2007) 908–912.

[30] G. Anitescu, A. Deshpande, L.L. Tavlarides, Integrated technology for supercriticalbiodiesel production and power cogeneration, Energy Fuel 22 (2008) 1391–1399.

[31] S.A. D' Ippolito, J.C. Yori, M.E. Iturria, C.L. Pieck, C.R. Vera, Analysis of a two-step,non-catalytic, supercritical biodiesel production process with heat recovery,Energy Fuels 21 (2007) 339–346.

[32] N. Cater, Making biodiesel in a microreactor, Tribol. Lubr. Technol. 62 (2006)15–24.

[33] J. Sun, J. Ju, L. Ji, L. Zhang, N. Xu, Synthesis of biodiesel in capillary microreactors,Ind. Eng. Chem. Res. 47 (2008) 1398–1403.

[34] G. Guan, K. Kusakabe, N. Sakurai, K. Moriyama, Continuous production of biodieselusing a microtube reactor, Chem. Eng. Trans. 14 (2008) 237–242.

[35] G. Guan, K. Kusakabe, K. Moriyama, N. Sakurai, Transesterification of sunflower oilwith methanol in a microtube reactor, Ind. Eng. Chem. Res. 48 (2009) 1357–1363.

[36] G. Guan, N. Sakurai, K. Kusakabe, Synthesis of biodiesel from sunflower oil at roomtemperature in the presence of various co-solvents, Chem. Eng. J. 146 (2009)302–306.

[37] I.Vieitez, C. Silva,G.R. Borges, F.C. Corazza, J.V.Oliveira,M.A.Grompone, I. Jachmanián,Continuous production of soybean biodiesel in supercritical ethanol–watermixtures,Energy Fuels 22 (2008) 2805–2809.

[38] C. Silva, Produção Contínua de Biodiesel por Transesterificação não catalítica deÓleo de Soja, Dr. Thesis, Universidade Estadual de Maringá, Brazil (in Portuguese),2009.

[39] Standard UNE-EN 14103: Determination of ester and linolenic acid methyl estercontents, issued by Asociación Española de Normalización y Certificación, Madrid,2003.

[40] R.E. Walker (Ed.), Official Methods and Recommended Practices of the AmericanOil Chemists' Society (Method AOCS Ce 2-66), 4th Ed., American Oil Chemists'Society, Champaign, 1990.

[41] I. Vieitez, C. Silva, I. Alkimim, G.R. Borges, F.C. Corazza, J.V. Oliveira, M.A.Grompone, I. Jachmanián, Effect of temperature on the continuous synthesis ofsoybean esters under supercritical ethanol, Energy Fuels 23 (2009) 558–563.

[42] Standard UNE-EN 14105: Determination of free and total glycerol and mono-, di-,triglyceride contents, issued by Asociación Española de Normalización yCertificación, Madrid, 2003.

[43] M.M. Gui, K.T. Lee, S. Bhatia, Supercritical ethanol technology for the production ofbiodiesel: process optimization studies, J. Supercrit. Fluids 49 (2009) 286–292.

[44] K.T. Tan, K.T. Lee, A.R. Mohamed, Production of FAME by palm oil transesterifica-tion via supercritical methanol technology, Biomass Bioenergy 33 (2009)1096–1099.

[45] N. Aimaretti, D.I. Manuale, V.M. Mazzieri, C.R. Vera, C. Yori, Batch study of glyceroldecomposition in one-stage supercritical production of biodiesel, Energy Fuel 23(2009) 1076–1080.

[46] K. Bunyakiat, S. Makmee, R. Sawangkeaw, S. Ngamprasertsith, Continuousproduction of biodiesel via transesterification from vegetable oils in supercriticalmethanol, Energy Fuels 20 (2006) 812–817.

[47] M.N. Varma, G. Madras, Synthesis of biodiesel from castor oil and linseed oil insupercritical fluids, Energy Fuels 37 (2007) 1–6.

[48] C.Wang, J. Zhou,W. Chen,W.Wang, Y. Wu, J. Zhang, R. Chi, W. Ying, Effect of weakacids as a catalyst on the transesterification of soybean oil in supercriticalmethanol, Energy Fuels 22 (2008) 3479–3483.

[49] H. Imahara, J. Xin, S. Saka, Effect of CO2/N2 addition to supercritical methanol onreactivities and fuel qualities in biodiesel production, Fuel 88 (2009) 1329–1332.

[50] C. Bertoldi, C. Silva, J.P. Bernardon, M.L. Corazza, L. Cardozo Filho, J.V. Oliveira, F.C.Corazza, Continuous production of biodiesel from soybean oil in supercriticalethanol and carbon dioxide as co-solvent, Energy Fuels 23 (2009) 5165–5172.