Microchemical Journal 109 (2013) 4650

Contents lists available at SciVerse ScienceDirect

Microchemical Journal
journal homepage: www.elsevier.com/locate/microc

Monitoring the conversion of soybean oil to methyl or ethyl esters using the refractive index with correlation gas chromatography
Regina. C. R. Santos, Rmulo B. Vieira, Antoninho Valentini
Langmuir Laboratrio de Adsoro e catlise, Departamento de Qumica Analtica e Fsico-Qumica, Universidade Federal do Cear, Fortaleza, Cear, Brazil CEP: 60.451.970

a r t i c l e

i n f o

a b s t r a c t
A simple method (refractive index) was applied to monitor the progress and the end point of the transesterication reaction of soybean oil to biodiesel (methyl or ethyl esters). Additionally, the same method may be used to determine the methyl or ethyl ester content during the transesterication reaction progress. To do so, blends of biodiesel and soybean oil were prepared at different wt.% to obtain a simple linear correlation with the refractive index and a correlation coefcient (R2) of 0.9997 and 0.9996 for the FAMEs and FAEEs, respectively. The transesterication process of soybean oil with methanol and ethanol was performed to determine how the refractive index properties change due to the ratio of conversion. It was concluded that in the reaction kinetics of the methanolysis reaction, the efciency was over 90% in 8 h. Compared with existing chromatographic techniques, the refractive index method for monitoring the transesterication of vegetable oils presented good results. Additionally, the method was rapid, inexpensive and especially suitable for process control applications. 2012 Elsevier B.V. All rights reserved.

Article history: Received 19 November 2011 Received in revised form 16 April 2012 Accepted 2 May 2012 Available online 9 May 2012 Keywords: Biodiesel Transesterication Soybean oil Refractive index Gas chromatography

1. Introduction With the increasing energy crisis due to fossil fuel depletion and environmental degradation, considerable effort has been devoted to the development of alternative fuels [1]. One of the most promising alternative diesel fuels is biodiesel, which is derived from the transesterication reaction of vegetable oils or animal fats with alcohol of lower molecular weights, such as methanol or ethanol, using homogeneous or heterogeneous catalysts (acid, base, or enzyme) [24]. The primary advantages of using biodiesel are that it is biodegradable and nontoxic, can be used without modifying existing engines, and produces less harmful gas emissions such as SOx, CO, unburned hydrocarbons, and particulate material [5,6]. However, incomplete transesterication and/or insufcient purication that results in even a small amount of the original unconverted oil compounds getting into the nal methyl or ethyl ester product can cause severe operational problems, such as engine deposits, which can increase the production of hazardous emissions [13,7]. Hence, monitoring the conversion of vegetable oils to methyl or ethyl esters is necessary to assess the quality of biofuels. Different methods have been used for determining or verifying the concentration of biodiesel obtained by the transesterication of vegetable oils. These include techniques such as 1H NMR spectroscopy [810], infrared (IR) spectroscopy [10,11], high-performance liquid

Corresponding author. Tel.: + 55 85 3366 9951; fax: + 55 85 3366 9982. E-mail address: valent@ufc.br (A. Valentini). 0026-265X/$ see front matter 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.microc.2012.05.001

chromatography (HPLC) [10,12] and gas chromatography [13,14]. However, gas chromatography (GC) [14,15] is the analytical method that is typically used for evaluating FAME conversion and quality according to European standard specications based on the EN in 14103. Gas chromatography (GC) is a very sensitive analytical method that measures by-products such as monoglyceride (MG), diglyceride (DG) and unreacted triglyceride (TG), in addition to fatty acid methyl ester (FAME). Nevertheless, this method has some drawbacks, such as requiring very accurate sample preparation. For example, prior to analysis by trimethylsilylation of the free hydroxyl groups in MGs and DGs, the sample must be derivatised. Furthermore, the analysis is time-consuming, and the method requires expensive instruments with highly trained personnel. Because of these requirements, on-line application in a transesterication factory would be very difcult [1315]. Thus, low cost and rapid analytical methods are desirable. An alternative non-destructive, rapid analytical technique that does not require sample pre-treatment involves measuring refractive indices; this method has received little attention for use in monitoring the quality of biodiesel [16,17]. Because the possible components of the reaction mixture, i.e., MG, DG, TG, and methyl or ethyl esters, have different refractive indices, which change from base oil to biodiesel properties, they can be correlated to the mixture composition to give an indication of the conversion rate of the transesterication reaction. In the present study, we investigated the use of refractive index measurements for monitoring a transesterication reaction process. This method is simple, rapid, and inexpensive. Thus, it is especially

R.C.R. Santos et al. / Microchemical Journal 109 (2013) 4650

47

suitable for process control applications. As a means of correlating and cross-checking the results with another analytical method, gas chromatography (GC) was selected. Thus, we performed a series of transesterication reactions using rened soybean oil with methanol and ethanol, and the methyl or ethyl ester content of the samples were determined by the refractive index and gas chromatography. 2. Experimental 2.1. Materials Rened bleached deodorised soybean oil was purchased from local stores, and its physicochemical properties are illustrated in Table 1. The reagents used during synthesis and purication procedures were methanol (99.9%), anhydrous ethanol (99.8%), sodium hydroxide (97%) pellets, and hydrochloric acid (37%). Methyl and ethyl laurate (methyl or ethyl dodecanoate, 98.0% and 95.5%, respectively) were used as an internal standard for GC analysis. 2.2. Synthesis of the methylic and ethylic biodiesel and soybean oil blends for the standard correlation To obtain a correlative equation among the refractive indices and the ester content, mixtures of soybean oil with methylic or ethylic biodiesel at different proportions were prepared, and the refractive index was measured in each sample. The biodiesel used in these mixtures was synthesised by repeating an alkaline transesterication reaction twice, using 0.6% (w/w) or 0.8% (w/w) sodium hydroxide as a catalyst for methanol (FAME fatty acid methyl ester) or ethanol (FAEE fatty acid ethyl ester), respectively. The alcohols were used in excess at 100% vol. The reaction temperatures were 25 C and 60 C for the methylic and ethylic routes, respectively, with a reaction time of 1 h. The transesterication reaction was performed in a 250mL three-neck at-bottom ask, equipped with a reux condenser, a thermometer and magnetic stirring. The reactor was initially charged using 50 g of soybean oil and heated to the reaction temperature. Subsequently, the above mentioned amount of catalyst was dissolved in alcohol and added to the reactor, marking the beginning of the reaction. After the end of the reaction period (1 h), the glycerol phase was discarded, and the same amount of alcohol and catalyst used in the previous step was added, and the procedure was repeated. Finally, the resulting biodiesel sample was washed with distilled water at least three times to remove the remaining catalyst and glycerol; the water volume corresponded to 10 wt.% of the biodiesel. However, for the rst washing, an HCl (0.1 N) solution was used to neutralise the alkali catalyst. The product was heated to 105 C under vacuum for the subsequent analyses. The main properties of the biodiesel were evaluated and are shown in Table 1. In the next step, soybean oil and biodiesel mixtures were prepared in different weight proportions, as follows: B0: 100% soybean oil, B15: 15% biodiesel + 75% oil, B30: 30% biodiesel + 60% oil, B45: 45% biodiesel + 55% oil, B60: 60% biodiesel + 40% oil, B70: 70% biodiesel + 30%

oil, B85: 85% biodiesel + 15% oil and B100: 100% biodiesel. The refractive index of these mixtures was measured, and the correlative equation for the refractive index with the ester content was tted and evaluated. The blends were prepared by weight because the weight fraction does not change with temperature. 2.3. Product analysis (GC analysis for fatty acid esters) Biodiesel analyses (fatty acid methyl esters, FAME and fatty acid ethyl esters, FAEE) were performed using a gas chromatograph equipped with a ame-ionisation detector and a capillary non-polar column measuring 30 m in length, 0.25 mm in internal diameter, and 0.25 m in lm thicknesses. The column temperature program was as follows: initial temperature of 135 C, hold for 3 min, and ramp at 15 C/min up to 230 C, hold at 20 min. Nitrogen was used as the carrier gas at 1.0 mL/min. The temperatures of the injector and detector were 250 C. The preparation of the sample for analysis and quantication was performed following standard method EN 14103 [18] using methyl or ethyl laurate as an internal standard. The injection was performed in split mode with a split ratio of 80:1 and sample size of 1.0 L. The biodiesel yield (per cent FAMEs or per cent FAEEs) was calculated using Eq. (1): FAMEs or FAEEs Aester C EI V EI f ester laurate =AEI mx P EI 1

where, Aester is the peak area of FAMEs or FAEEs, CEI and VEI are the concentration and volume of the methyl or ethyl laurate solution, respectively, f methyl or ethyl laurate is the response factor, AEI and PEI are the peak area and purity (wt.%), respectively, of the internal standard and m is the mass of the sample. 2.4. Physicochemical properties' measurement of soybean oil and biodiesel The refractive index was measured to an accuracy of 10 4 with an ABBE refractometer at 40 C, which was thermostatically controlled by bath to maintain the temperature, in keeping with ISO 6320. The measurements were conducted three times for each sample, and the results were averaged. Physicochemical analyses of the soybean oil and biodiesel were performed according to ASTM methods D4052 and D445 for density and kinematic viscosity, respectively, and AOCS Cd 3d-63 for the acid value [1921]. The values obtained for these properties were compared to the European specications (EN 14214), for which the accepted values fall between 0.860 and 0.900 g/cm 3 for density, 3.5 and 5.5 mm 2/s for viscosity, and 0.5 mgKOH/g for the maximum acid value. 2.5. Synthesis of the methylic and ethylic biodiesel for the evaluation of the method To obtain real data from the soybean oil transesterication reaction progression, a new process of transesterication to methyl esters and ethyl ester was performed; however, a heterogeneous catalyst (dolomite) was employed [22]. It is known that a homogeneous process is faster than a heterogeneous process [23]; thus, this method makes it possible to obtain mixtures with compositions that are very different from that of the standard in Section 2.2. Methyl esters were produced at different reaction times (1, 2, 4, 8, 16 and 24 h), at 50 C and with a catalyst amount of 0.6 wt.% and a methanol/oil molar ratio of 6. Ethyl ester synthesis was performed at different temperatures (50, 55, 60, 70 and 80 C), maintaining the reaction time at 24 h, with a catalyst amount of 0.8 wt.% and an ethanol/oil molar ratio of 6. These new samples were monitored by gas chromatography according to Eq. (1) and compared to the calculated

Table 1 Soybean oil and biodiesel properties. Property Acid value Kinematic viscosity at 40 C Density at 20 C Refractive index at 40 C Fatty acid esters content Units Soybean oil FAME 0.48 4.2 0.881 1.4498 96.54 FAEE 0.29 4.5 European spec. 0.5 max. 3.55.5

mg KOH/g 0.099 mm2/s 30.1 g/cm3 0.9237 1.4680

0.876 0.860.9 1.4480 98.2 96.5

% mass

48

R.C.R. Santos et al. / Microchemical Journal 109 (2013) 4650

data using the correlations from measurements of the refractive indices. 3. Results and discussion 3.1. Physicochemical characterisation of pure soybean oil and biodiesel Some of the main properties of the soybean oil and the biodiesel synthesised by the homogeneous method are listed in Table 1, which also lists the comparisons of the obtained soybean FAME and FAEE with the European specication limits. The properties of the obtained biodiesel, in general, are similar. However, the triacylglyceride conversion into methyl and ethyl esters through the transesterication process, as evaluated by gas chromatography, revealed a higher ester content than soybean FAEE. The acid numbers for both samples meet the standard limit, indicating that the free fatty acid content will not cause operational problems, such as corrosion and/or pump plugging from corrosion-induced deposit formation. With regard to the density and viscosity, both are within the specication limits, indicating that the sample can be used directly in a diesel engine. A chromatogram of the FAMEs and FAEE obtained from the soybean oil transesterication reaction is shown in Fig. 1. The analysis of the main compounds in the biodiesel was identied by CG/MS. It is clear that the biodiesel obtained in the experiment primarily contains palmitate, oleate, linoleate, and linolenate fatty acid esters (Fig. 1). 3.2. Refractive index and ester content correlation The refractive index of a liquid is easily measured and does not require expensive and complicated equipment. To correlate the wt.% of the total esters of soybean oil and biodiesel mixtures studied in this work to their refractive indices, regression analysis based on the experimental results was performed. Fig. 2 (a and b) shows the variation in the physical properties as a function of weight per cent of total methyl and ethyl esters in the blends. As can be observed there was a linear correlation between the FAME and FAEE ratios and the refractive index, which decreased with the increase in the

Fig. 2. Refractive index of the biodiesel blends at 40 C versus wt.% total methyl ester (a) and (b) ethyl ester.

ester content of the samples. Such a linear correlation further conrms the reliability of the method. For the correlations with wt.% of the total esterrefractive index, Eqs. (2) and (3) for the methyl ester + soybean oil and ethyl ester + soybean oil mixtures, respectively, and their R-squared value were obtained. Y 1:90735104 X 1:4679 R2 0:9997 2 3

Y 1:99958104 X 1:4675 R2 0:9996

Fig. 1. The chromatographic prole of the soybean oil fatty acid methyl and ethyl esters. 1 = Internal standard (Lauric acid ester); 2 = Hexadecanoic acid ester; 3 = 9,12-Octadecadienoic acid ester; 4 = 9-Octadecenoic acid ester; 5 = Stearic acid ester.

In the equation, Y is the refractive index as a function of X, which is the weight per cent total acid fatty ester (FAME or FAEE) in the sample. Thus, due to the good R-squared value, the suggested correlation could be used to determine the conversion to methyl or ethyl esters in the transesterication process by measuring the refractive index. However, some real data points are needed to evaluate predictions using the above correlations. Therefore, as described in Section 2.5, a new series of methyl and ethyl esters were synthesised. Fig. 3 shows the methyl ester production from soybean oil, which was determined by gas chromatography (CG) and the refractive index (RI). It is important to note the excellent correlation obtained for the different methods. Fig. 3 shows that both techniques reveal a soybean oil conversion of approximately 90% in the rst 8 h of reaction time. In fact, after this point, the reactions continue, but the conversion rate is very low, indicating that reaction equilibrium was

R.C.R. Santos et al. / Microchemical Journal 109 (2013) 4650

49

Fig. 3. The conversion of soybean oil to methyl esters at various reaction times determined by gas chromatography (CG) and the refractive index (RI). Reaction conditions: catalyst amount, 0.6 wt.%; methanol/oil molar ratio, 6; temperature, 50 C.

Fig. 4. The conversion of soybean oil to ethyl esters at various reaction temperatures, as determined by gas chromatography (GC) and the refractive index. Reaction conditions: catalyst amount, 0.8 wt.%; ethanol/oil molar ratio, 6; reaction time, 24 h.

almost reached. In the reaction time from 1 to 8 h, a modest divergence in the methyl ester content is calculated by the different methods. However, during this reaction time (from 1 to 8 h), the refractive index undergoes a signicant change, primarily due to the monoglyceride and diglyceride content. While the standard mixture (Fig. 2), is composed primarily of methyl ester and triglycerides, the reaction mixture (Fig. 3) is composed of methyl ester, triglycerides and a substantial ratio of monoglycerides and diglycerides, which have different refractive indices [16,24]. The presence of mono- and diglycerides is indicated by the per cent relative error for each data point reported in Table 2. The higher relative error observed for the samples from the rst 8 h of reaction time is due to the difculties in producing a standard mixture with a composition that is very close to the real mixture, which contains mono and diglycerides in addition to triglycerides and the esters. A very low relative error was observed for the sample, which reached a good conversion. Fig. 4 presents the effect of reaction temperature (50, 55, 60, 70 and 80 C) on the soybean oil conversion for FAEE, for a xed reaction time of 24 h. As expected, an increase in temperature promotes higher conversion. Thus, the equilibrium position shifted toward the

Table 2 FAME or FAEE soybean oil conversions measured by gas chromatographya and calculated brefractive index properties and their relative per cent error. FAMEs Reaction time (h) Calc. Eq. (1) GC a Calc. Eq. (2) RI b Relative error % 1 13.50 14.73 9.111 FAEEs Temperature (C) Calc. Eq. (1) GC a Calc. Eq. (3) RI b Relative error %
a b

ester as the temperature increased, in agreement with data from the literature [2527]. Although the reaction kinetics of ethanolysis are slower, if they are compared to methanolysis, there is no signicant difference regarding the accuracy of the FAEEs calculated by the refractive index using Eq. (3) and gas chromatography using Eq. (1); their values change very little (Fig. 4). The low relative error observed for FAEE production may suggest that the ratio of mono and diglycerides is lower than that of FAME production. In other words, the ethanolysis of mono and diglycerides is much faster than the ethanolysis of triglycerides. Aside from the kinetics steps and each rate value, according to Table 2, the FAME and FAEE relative errors are less than 10% and 2%, respectively, and the maximum mean relative errors were 4.339% and 0.847% for FAME and FAEE, respectively. Although there are some deviations in these data and correlations, the differences may occur primarily when the equilibrium of the reaction has not yet been achieved, at which point the refractive index changes signicantly during a short time as the composition of the reaction mixture changes; however, in the second part of the reaction, at equilibrium, the composition is nearly constant and very similar to the standard mixtures [24]. Therefore, the relative errors indicate that the equations obtained from the proposed analytical method, which is rapid and low-cost, are adequate for monitoring the progress of the soybean oil transesterication reaction. These methods were developed using soybean oil instead of other vegetable oils. Specic calibration curves for each vegetable oil should be prepared because of their different compositions. 4. Conclusions Equations to correlate the wt.% total methyl or ethyl ester with the refractive index for biodiesel plus soybean oil mixtures were proposed. As observed in this study, the refractive index decreased linearly with the increase of methyl or ethyl ester wt.% from soybean oil. Thus, a very good linear relationship was obtained with an R2 of 0.9997 (FAMEs) and 0.9996 (FAEEs) for the blends. Comparisons between the experimental data obtained by gas chromatography and the values predicted by the proposed correlations showed that the per cent mean relative error observed was higher for FAME than FAEE. Such relative error may occur because the state of reaction equilibrium is not reached. However, in general, there is a good agreement between the gas chromatography data and the values predicted by the proposed equations, suggesting that this methodology can be very useful for the

2 28.50 31.05 8.947

4 75.50 78.17 3.536

8 92.80 89.18 3.901

16 93.70 93.89 0.203

24 94.10 94.42 0.340

Mean R.E. %

4.339

50 13.30 13.20 0.752

55 16.50 16.70 1.212

60 91.53 91.22 0.338

70 95.70 94.22 1.546

80 97.10 96.72 0.391

Mean R.E. %

0.847

From Eq. (1) according to EN 14103 based on the different reaction conditions. From Eqs. (2) and (3) based on blend data.

50

R.C.R. Santos et al. / Microchemical Journal 109 (2013) 4650 [12] M. Holcapek, P. Jandera, Jan Fischer, B. Prokes, Analytical monitoring of the production of biodiesel by high performance liquid chromatography with various detection methods, J. Chromatogr. A 858 (1999) 1331. [13] C. Plank, E. Lorbeer, Simultaneous determination of glycerol, and mono-, diand triglycerides in vegetable oil methyl esters by capillary gas chromatography, J. Chromatogr. A 697 (1995) 461468. [14] B. Freedman, W.F. Kwolek, E.H. Pryde, Quantitation in the analysis of transesteried soybean oil by capillary gas chromatography, J. Am. Oil Chem. Soc. 63 (1986) 13701374. [15] G. Knothe, Analyzing biodiesel: standards and other methods, J. Am. Oil Chem. Soc. 83 (2006) 823833. [16] W. Xie, H. Li, Hydroxyl content and refractive index determinations on transesteried soybean oil, J. Am. Oil Chem. Soc. 83 (2006) 869872. [17] I. Nita, S. Geacai, O. Iulian, Measurements and correlations of physico-chemical properties to composition of pseudo-binary mixtures with biodiesel, Renew. Energy 36 (2011) 34173423. [18] EN14103, Fat and oil derivativesfatty acid methyl esters (FAME) determination of ester and linolenic acid methyl ester contents, 2003. [19] ASTM D4052, Density and Relative Density of Liquids. Digital Density Meter, 2001. [20] ASTM D445, Kinematic Viscosity of Transparent and Opaque Liquids (and the Calculation of Dynamic Viscosity), 2001. [21] AOCS Cd 3d-63. Acid Value. In: Firestone, D. (Eds.), Ofcial Methods and Recommended Practices of the American Oil Chemists' Society, Champaign, IL, USA, 2009. [22] O. Ilgen, Dolomite as a heterogeneous catalyst for transesterication of canola oil, Fuel Process. Technol. 92 (2011) 452455. [23] F. Qiu, Y. Li, D. Yang, X. Li, P. Sun, Heterogeneous solid base nanocatalyst: preparation, characterization and application in biodiesel production, Bioresour. Technol. 102 (2011) 41504156. [24] N. Ellis, F. Guan, T. Chen, C. Poon, Monitoring biodiesel production (transesterication) using in situ viscometer, Chem. Eng. J. 138 (2008) 200206. [25] O.S. Stamenkovic, A.V. Velickovic, V.B. Veljkovic, The production of biodiesel from vegetable oils by ethanolysis: current state and perspectives, Fuel 90 (2011) 31413155. [26] S.M.P. Meneghetti, M.R. Meneghetti, C.R. Wolf, E.C. Silva, G.E.S. Lima, L.L. Silva, T.M. Serra, F. Cauduro, L.G. Oliveira, Biodiesel from castor oil: a comparison of ethanolysis versus methanolysis, Energy Fuel 20 (2006) 22622265. [27] W. Zhou, S.K. Konar, D.G.B. Boocock, Ethyl esters from the single-phase base-catalyzed ethanolysis of vegetable oils, J. Am. Oil Chem. Soc. 80 (2003) 367371.

rapid and low-cost monitoring of the conversion of soybean oil to methyl or ethyl ester in the biodiesel production process. Acknowledgements The authors acknowledge the Universidade Federal do Cear (UFC) and the Brazilian research-funding support agencies CNPq. References
[1] M.A. Fazal, A.S.M.A. Haseeb, H.H. Masjuki, Biodiesel feasibility study: an evaluation of material compatibility; performance; emission and engine durability, Renew. Sustain. Energ. Rev. 15 (2011) 13141324. [2] I.M. Atadashi, M.K. Aroua, A. Abdul Aziz, High quality biodiesel and its diesel engine application: a review, Renew. Sustain. Energ. Rev. 14 (2010) 19992008. [3] L.C. Meher, D.V. Sagar, S.N. Naik, Technical aspects of biodiesel production by transesterication a review, Renew. Sustain. Energ. Rev. 10 (2006) 248268. [4] J. Janaun, N. Ellis, Perspectives on biodiesel as a sustainable fuel, Renew. Sustain. Energ. Rev. 14 (2010) 13121320. [5] J. Qian, F. Wang, S. Liu, Z. Yun, In situ alkaline transesterication of cottonseed oil for production of biodiesel and nontoxic cottonseed meal, Bioresour. Technol. 99 (2008) 90099012. [6] F. Qiu, Y. Li, D. Yang, X. Li, P. Sun, Biodiesel production from mixed soybean oil and rapeseed oil, Appl. Energy 88 (2011) 20502055. [7] N. Maheshwari, C. Balaji, A. Ramesh, A nonlinear regression based multi-objective optimization of parameters based on experimental data from an IC engine fueled with biodiesel blends, Biomass Bioenergy 35 (2011) 21712183. [8] G. Knothe, Determining the blend level of mixtures of biodiesel with conventional diesel fuel by ber-optic near-infrared spectroscopy and 1H nuclear magnetic resonance spectroscopy, J. Am. Oil Chem. Soc. 78 (2001) 10251028. [9] M. Morgenstern, J. Cline, S. Meyer, S. Cataldo, Determination of the kinetics of biodiesel production using proton nuclear magnetic resonance spectroscopy (1H NMR), Energy Fuel 20 (2006) 13501353. [10] G. Knothe, Analytical methods used in the production and fuel quality assessment of biodiesel, Am. Soc. Agric. Eng. 44 (2001) 193200. [11] M.F. Pimentel, G.M.G.S. Ribeiro, R.S. da Cruz, L. Stragevitch, J.G.A. Pacheco Filho, L.S.G. Teixeira, Determination of biodiesel content when blended with mineral diesel fuel using infrared spectroscopy and multivariate calibration, Microchem. J. 82 (2006) 201206.