Research Article Received: 29 July 2013,
Revised: 20 November 2013,
Accepted: 26 November 2013
Published online in Wiley Online Library: 23 January 2014
(wileyonlinelibrary.com) DOI 10.1002/pca.2499
Rapid Separation of Free Fatty Acids in Vegetable Oils by Capillary Zone Electrophoresis Renata Takabayashi Sato,a Renata de Jesus Coelho Castro,a Patrícia Mendonça de Castro Barraa and Marcone Augusto Leal de Oliveiraa* ABSTRACT: Introduction – Olive oil is a very important product to human health since it inhibits formation of free radicals, tumour growth, lesions and inflammatory substances. High concentrations of free fatty acids in olive oils results in lipid deterioration due to oxidative or hydrolytic rancidity. Objective – To optimise an alternative capillary zone electrophoresis methodology, under ultraviolet indirect detection and to determine free fatty acids in edible vegetable oils without derivatisation steps in sample preparation. Methods – The condition used consisted of 15 mM NaH2PO4–Na2HPO4 at pH ~6.86, 4.0 mM of sodium dodecybenzenesulphonate, 8.3 mM of Brij 35®, 45% v/v of acetonitrile and 2.1% of 1-octanol, injection at 12.0 mbar of pressure for 4 s, +19 kV of applied voltage and indirect detection at 224 nm, within an analysis time of 11 min. Results – The capillary zone electrophoresis method was successfully applied to determination of free fatty acids in extra virgin olive oil, virgin olive oil and soybean oil samples. The comparison with the official volumetric titration method showed no significant difference within the 95% confidence interval. Conclusion – The main advantage to the proposed method is the possibility to observe the individual amount of the free fatty acids, which would be useful for researchers interested in studying the effect of the free fatty acids profile on oxidative process in food. Copyright © 2014 John Wiley & Sons, Ltd. Keywords: Capillary zone electrophoresis; free fatty acids; vegetable oil
Introduction
Phytochem. Anal. 2014, 25, 241–246
* Correspondence to: M. A. L. de Oliveira, Departamento de Química, Instituto de Ciências Exatas, Universidade Federal de Juiz de Fora, Cidade Universitária CEP 36036-330, Juiz de Fora, MG, Brazil. E-mail: [email protected] a
Departamento de Química, Instituto de Ciências Exatas, Universidade Federal de Juiz de Fora, Cidade Universitária CEP 36036-330, Juiz de Fora, MG, Brazil
Copyright © 2014 John Wiley & Sons, Ltd.
241
Olive oil is a very important product to human health because it inhibits formation of free radicals, tumour growth, lesions and inflammatory substances, improves the lipidic profile and prevents ageing of cells (de Oliveira et al., 2008). The best quality olive oil is the extra virgin, which is prepared from the first cold pressing of the olives and has a higher commercial price compared with the other oils. In many countries, rigid rules have been created for the consumer’s protection, taking into account different analytical possibilities to identify the authenticity and quality of the olive oil. One of the parameters evaluated for olive oil quality is the free acidity, which is expressed through the oleic acid (C18:1 9c) percentage (Table 1) (European Commission Regulation, 1991). The hydrolytic rancidity is produced from the hydrolysis of the ester bond by lipase or by moisture, resulting in the production of free fatty acids (FFA). The FFA formed are responsible for the taste and unpleasant odour, especially in fats such as butter, which has a big amount of low mass molecular fatty acid (FA). However, in fats containing non-volatile FA, such as in edible vegetable oils, the characteristic taste and odour do not appear together with the deterioration (Osawa et al., 2006). High concentration of FFA in edible oils results in lipid deterioration, which constitutes one of the most important problems in the olive oil industry. Studies performed with rats have shown that the presence of rancid food in diet might contribute to partial inactivation or destruction of vitamins A, D, E and K, carotene, tocopherols,
ascorbic acid and essentials fatty acids; in addition, it changes the digestive system operation (Greenberg and Frazer, 1953). On the other hand, the FFA accumulation in plants, for example, might indicate that they were exposed to environmental stress such as freezing or desiccation (Barclay and Mckersie, 1994). Moreover, Barclay and Mckersie’s (1994) work has shown that the determination of individual FFA is also important because the free palmitic acid may reduce aldehyde production, while free linolenic acid increases the aldehyde and decreases the fluorescent product simultaneously. Within this context, the development and optimisation of fast and reliable alternative analytical methods for oil acidity determination is very necessary. It is important to highlight that such methods can be useful to help in quality control or in the monitoring of acidity in stored olives, the extracted olive oil or even to attest the quality of the product available for consumption on the market shelf. The classic method for olive oil acidity determination is alkaline volumetric titration (AVT) (European Commission Regulation,
R. T. Sato et al. Table 1. Vegetable oil classification according to acidity Product Extra virgin olive oil Virgin olive oil Common virgin olive oil Refined olive oil Olive oil Lampante olive oil
Acidity, expressed in oleic acid (g/100 g) 0.8 2.0 3.3 0.3 1.0 >3.3
1991; Mariotti and Mascini, 2001). However, other analytical approaches have been used, such as flow injection analysis in automated systems (Nouros et al., 1997; Mariotti and Mascini, 2001), flow-reversal injection liquid–liquid extraction (Zhi et al., 1996), voltammetric reduction of quinones (Takamura et al., 1999), infrared and near infrared spectroscopy (Lai et al., 1995; Cayuela et al., 2009), Raman spectroscopy (Aparicio and Baeten, 1998), flow injection analysis (FIA) (Makahleh and Saad, 2011) and gas chromatography (Lercker and Rodriguz-Estrada, 2000; Kanya et al., 2007). Since the 1990s, capillary electrophoresis (CE), a separation technique based on differentiatial migration of neutral, ionic and ionisable compounds by application of an electrical field in a fused silica capillary tube filled with a convenient electrolyte solution has been successfully applied as an alternative method to FA analysis in different matrixes (Balesteros et al., 2007; De Castro et al., 2010; Porto et al., 2011; Barra et al., 2012, 2013; Castro et al., 2013). This present study takes into account work performed by Balesteros et al. (2007), which was the first literature account of olive oil acidity analysis by capillary zone electrophoresis (CZE). However, at the time, the methodology proposed was not so attractive for users with little experience of the CE technique, due to the necessity of removing the external polyimide coating in the ends of the capillary in order to avoid adsorption problems. This phenomenon was observed after detailed investigation in which the coating became partially dissolved in the electrolyte and the polymer gradually adhered to the capillary internal wall, disturbing the separation profile significantly and consequently decreasing the lifetime of the capillary. Furthermore, the method was applied only to extra virgin olive oil (EVOO) acidity determination. Within this context, the main goal of the present work was to demonstrate the potential of CE for FFA analysis in vegetable oils, such as EVOO, virgin olive oil (VOO) and soybean oil (SO) samples. In addition, because the method makes it possible to determine individual FFA, the oleic acid may be used as a marker to indicate possible fraudulant EVOO and, consequently, would be useful in product quality control. The proposed method therefore can be proposed as an attractive alternative to the official method by AVT (European Commission Regulation, 1991), which uses phenolphthalein as indicator, in terms of good sampling rate, economy of solvents and sample and reducing human intervention.
Materials and methods Chemicals and materials
242
All the reagents were of analytical grade and the water was purified by deionisation (Milli-Q system; Millipore, Bedford, MA, USA). The solvents methanol (MeOH), ethanol (EtOH) (Vetec, Rio de Janeiro, Brazil), acetonitrile (ACN) and 1-octanol (Merck,
wileyonlinelibrary.com/journal/pca
Rio de Janeiro, Brazil) of chromatographic grade were purchased. Polyoxyethylene 23 lauryl ether (Brij 35®) and sodium dodecylbenzenesulphonate (SDBS) were obtained from SigmaAldrich (St Louis, MO, USA), and sodium hydroxide (NaOH) was obtained from Synth (São Paulo, Brazil). Tridecanoic (C13:0), palmitic (C16:0), oleic (C18:1 9c) and linoleic (C18:2 9c, 12c) FA standards were purchased from Sigma-Aldrich. Individual FA stock solutions at a concentration of 20.0 mM were prepared by dissolving appropriate amounts of the above-mentioned standards in methanol and they were then stored in a freezer until analysis. A mixture of all the standards was prepared at concentration of 0.5 mM by the appropriate dilutions in MeOH. The aqueous Brij 35® stock solution was prepared by weighing and dissolving an amount corresponding to 50.0 mM in a 100.0 mL volumetric flask. A mass of NaOH corresponding to 1.0 M was weighed and dissolved in a 100.0 mL volumetric flask and the volume was filled with deionised water. An aqueous SDBS stock solution was prepared by weighing and dissolving a mass corresponding to 100.0 mM in a 100.0 mL volumetric flask. The aqueous buffer (pH ~6.86) stock solutions at concentrations of 100.0 mM were prepared from a mass of sodium phosphate monobasic (NaH2PO4) corresponding to 50.0 mM and 50.0 mM of sodium phosphate dibasic (Na2HPO4), which was weighed and dissolved in a 250.0 mL volumetric flask. Phosphate buffers and the Brij 35® stock solutions were kept in a freezer to prevent mould formation. The fresh working electrolyte solution was prepared by the appropriate dilutions of stocks and the incorporation of solvents. Instrumentation for the capillary electrophoresis system The experiments were conducted using a CE system (HP3d CE, Agilent Technologies, Palo Alto, CA, USA) equipped with a diodearray detector (DAD), with ultraviolet (UV) indirect detection, at 224 nm, a temperature control device (set at 25 °C), data acquisition and treatment software (HP ChemStation, Rev. A.06.01). Samples were hydrodynamically injected (12.0 mbar for 4 s), and the electrophoretic system was operated under normal polarity and constant voltage (+19 kV); manual integration was performed using peak baselines. For all the experiments, a fused-silica capillary tube with fluoropolymer (TSH) external coating was used (Polymicro Technologies, Phoenix, AZ, USA): 48.5 cm long (40 cm effective length), 75 μm internal diameter (ID) and 375 μm outside diameter (OD). In the present case, the TSH capillary avoids irreversible deleterious adsorption into the internal capillary wall, which causes intense problems in the separation performance, as demonstrated by Balesteros et al. (2007). Sample preparation Capillary Electrophoresis. The EVOO, VOO and SO samples were purchased in local market. A key point in the sample preparation step is the weighed mass: the weighed mass of each sample takes into account the separation resolution profile presented. Thus, the higher the expected acidity, the smaller should be the mass to carry the weight: EVOO mass weight was 0.75 g, VOO mass weight was 0.50 g and SO mass weighted was 1.50 g. The weighed samples were transferred to 5.0 mL volumetric flask and 0.50 mM of C13:0 was added, as internal standard (IS), and the volume was completed with ethanol at 60 °C. The
Copyright © 2014 John Wiley & Sons, Ltd.
Phytochem. Anal. 2014, 25, 241–246
Analysis of Free Fatty Acids in Vegetable Oil By Cze volumetric flask was shaken manually and after cooling the volume was completed, again, with ethanol at room temperature. It was shaken by vortex for 2 min and maintained for 2 min at rest. An aliquot of the ethanolic phase (top) was injected into the CE equipment. The analyses were performed in authentic duplicates. It is important to mention that dissolution sample tests using other solvents, such as methanol and isopropanol were performed. However, the results were not satisfactory, because with these solvents higher signal-to-noise ratios were observed in the electropherograms profiles when compared with the use of ethanol. Titration. The analyses were performed using the official cold volumetric titration method according to European Commission Regulation (1991). The samples were titrated with NaOH 0.1 M, phenolphthalein as an indicator and toluene:ethanol mixture (1:1 v/v) as the extractor reagent. The analyses were performed in authentic duplicates. The sample preparation procedure for CE method and the titration are illustrated in Fig. 1.
Capillary electrophoresis procedures Before use, new capillaries were conditioned by pressure flushing with 1.0 mM NaOH (40 min), deionised water (15 min) and electrolyte solution (15 min). The capillaries were regenerated between runs by washing with 1.0 M NaOH (2 min), deionised water (2 min) and fresh electrolyte solution (2 min). This conditioning procedure was critical to ensure peak area and migration time repeatability and to prevent solute adsorption to the capillary wall.
Statistical analysis The statistical tests such as homoscedasticity and normality were performed using IBM®SPSS® Statistics 14.0 for Windows software. The lack of fit analysis was performed in Microsoft Office® Excel software.
Results and discussion Background electrolyte Traditionally, the analysis of FFA by CZE takes into account the background electrolyte (BGE) added to organic solvent such as ACN in order to avoid micelle formation among the FFA; pH higher than 7.0 in order to promote the carboxyl group dissociation (because FFA has pKa ≈ 5.0, they are analysed in anionic form) under counter electroosmotic flow (EOF), catodic EOF and the use of chromophoric agent (SDBS) to promote indirect detection, because saturated FFA present low molar absorptivity in the UV range. Thus, the BGE consisted of 15 mM of NaH2PO4– Na2HPO4 at pH ~6.86, 4.0 mM SDBS, 8.3 mM Brij 35®, 45% v/v ACN and 2.1% 1-octanol. Thus, the output order for each FFA takes into account the charge and size relationship. In the other words, the FFA that has a low charge and size relationship, resists the EOF less and reaches the detection window first. Response factor calculation The proposal of the FFA quantification in edible vegetable oil samples was based on a statistical study including the response factor (Rf) calculation by using C13:0 as the internal standard (IS). In order to calculate Rf, a random experiment in authentic replicates using FFA standards of C16:0, C18:1 9c and C18:2 9c, 12c in concentrations of 0.2, 0.6, 1.0, 1.4, 1.8 and 2.2 mM and IS in fixed concentration of 0.5 mM, was performed by means of the leastsquares method (LSM). It is important to note that the correct practice to evaluate the fit of the model include removal of points in each level if necessary. Table 2 shows the values used in the regression model. Thus, after careful statistical evaluation (α = 0.05), that is, the assumptions estimation such as homoscedasticity test (Levene: different numbers of replicates into the same level or Cochran: same numbers of replicates into the same level), residue normality test (Shapiro–Wilk) and lack of fit test through a priori test hypothesis (ANOVA), recommended by the IUPAC (Danzer and Currie, 1998), described in equation (1), where mi is the number of measurements, p represents the calibration points, and m is the product of p and mi, the regression model was implemented. p X
F calc ¼
S2y;x S2y
¼
mi ðy i y^i Þ2 =ðp 2Þ
i¼1
p X mi X
y ij y i
2
(1) =ðm pÞ
i¼1 j¼h
In the present case, the regression model diagnosis was adequate with no lack of fit because the value of Fcalc is lower than Fcrit for all FA in the 95% confidence interval for the concentration range studied and the Rf values found were 0.526, 0.622 and 0.604 for C18:1 9c, C16:0 and C18:2 9c, 12c, respectively, as summarised in Table 3. Thus, the quantification procedure involving the calculation of Rf is described by equation (2) AFFA AC13:0 ¼ Rf ½FFA ½C13 : 0
Phytochem. Anal. 2014, 25, 241–246
where AFFA is the area for each FFA, AC13:0 is the IS area, [FFA] is the concentration (in mM) for each FA and [C13:0] is the IS concentration fixed in 0.5 mM.
Copyright © 2014 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/pca
243
Figure 1. Schemes of sample preparations for the analysis of free fatty acids by capillary electrophoresis (CE) and volumetric titration.
(2)
R. T. Sato et al. Table 2. Values used for regression model implementation FFA
Signal first replicate
[FFA]/[IS]
C18:1 9c
1.2 2.0 2.8 3.6 4.4 0.4 1.2 2.0 2.8 0.4 2.0 2.8 3.6 4.4
C16:0
C18:2 9c, 12c
Signal second replicate
78.7 96.1 208.2 194.2 292.8 36.9 111.5 – 208.5 33.4 134.6 243.8 262.2 367.6
Signal thir replicate – – – – – – – – – 28.2 160.2 211.4 344.1 472.4
70.7 90.3 147.6 181.5 303.6 31.8 129.4 179.9 217.9 28.3 127.1 220.7 317.0 406.8
Table 3. Response factor calculation FFA
Slope
C18:1 9c C16:0 C18:2 9c, 12c
0.526 (±0.016) 0.622 (±0.024) 0.604 (±0.012)
Intercept
r
Fcalc
Fcrit
-0.143 (±0.048) -0.013 (±0.048) -0.078 (±0.037)
0.996 0.994 0.997
4.42 1.80 2.86
5.41a 6.59b 3.71c
a
Ftab(n1 = 3, n2 = 5). Ftab (n1 = 3, n2 = 4). c Ftab (n1 = 3, n2 = 10). Cochran test: Ccalc (C18:1 9c) = 0.32 , Ctab (C18:1 9c) = 0.93; Ccalc (C18:2 9c, 12c) = 0.63, Ctab (C18:2 9c, 12c) = 0.79; Levene test (p-value): (C16:0): 0.06; Shapiro–Wilk test (p-values): C18:1 9c = 0.636; C16:0 = 0.378; C18:2 9c, 12c = 0.041. b
As the regression model diagnosis was satisfactory, the slope can be used as a response factor in equation (2), and as long as the internal standard C13:0 at 0.5 mM is used, the concentration of FFA remains hidden. Thus, the acidity percentage in the sample was carried out through equation (3), obtained after rearranging equation (2). It is important to note that the CZE method allows the determination of individual FFA, but the % acidity is expressed in oleic acid percentage, because this is the majority FA in olive oils. However, to calculate the % acidity value, we must take into account the sum of all the FFA present in the sample areas. %acidity ¼
MMC18:0 ½C18 : 0V Rf AC13:0 m
X
Axi
100
(3)
Table 4. Comparison between capillary electrophoresis and volumetric titration methods Sample EVOO VOO SO
CE/%oleic
244
0.35 0.40 1.68 1.65 0.07 0.07
wileyonlinelibrary.com/journal/pca
acid
Titration/%oleic 0.36 0.35 1.20 1.20 0.07 0.07
acid
where Axi is the area for each FFA in samples (C18:1 9c, C16:0, C18:2 9c, 12c), AC13:0 is the IS area, [C13:0] is the IS fixed concentration of 0.5 mM, V is the volume in litres, m is the sample mass in milligrams, Rf is the response factor (fitted model slope) and MMC18:1 9c is the molecular mass for C18:1 9c (282.5 g/mol). This approach is very interesting, because Rf is calculated just once for each FFA in preliminary experiments (under controlled operational conditions) with standards. Thus, after rigorous statistical evaluation of the regression models through the homoscedasticity, residues and linearity test evaluation through the lack of fit test (ANOVA), which did not present significant evidence for the considered interval (in the present case all values obtained in statistical tests were considered acceptable within the 95% and 96% confidence intervals), the obtained slopes can be used as Rf for the sample quantification. On other words, if some analyses were performed using the same IS, in the same concentration of the preliminary experiment and using the same operational conditions (capillary dimensions, wavelength, cartridge temperature, etc.) and there were no violations (or meaningful evidence) in the statistic tests, the calculated Rf for each FFA can be used for different samples without the necessity to perform experiments to obtain the Rf every time.
Limit of detection and limit of quantification Calculation of the limits of detection (LOD) and quantification (LOQ) can be undertaken in three different ways: (i) standard
Copyright © 2014 John Wiley & Sons, Ltd.
Phytochem. Anal. 2014, 25, 241–246
Analysis of Free Fatty Acids in Vegetable Oil By Cze
Figure 2. Sample electropherograms. Peaks: (1) C18:1 9c; (2) C16:0; (3) C18:2 9c, 12c; (IS) C13:0. Electrolyte composition: 15.0 mM of sodium phosphate buffer (pH ~6.8); 8.3 mM of Brij 35®; 4.0 mM of SDBS; 2.1% of 1-octanol and 45% of ACN. Operational conditions: 4 s injection at 12.0 mbar pressure, +19 kV applied voltage, indirect detection at 224 nm and TSH capillary 40 cm effective length.
deviation of the response and the slope, (ii) signal-to-noise approach and (iii) visual evaluation. In the present case, the signal-to-noise approach was used. This approach can be applied only to analytical procedures that exhibit baseline noise, which the signal-noise relation through the standard deviation calculation of the baseline (noise) and the height of the peak (signal) of the analytes (ICH Expert Working Group, 2005). The LOD and LOQ were calculated within the concentration corresponding to a signal-noise relationship equal to 3 (equation (4)) and 10 (equation (5)), respectively, where Sb is the standard deviation of the baseline, CS is the concentration of analyte, Hmax is the maximum peak height and Hmin is the baseline level. LOD ¼
3 Sb C S H max H min
(4)
LOQ ¼
10 Sb C S H max H min
(5)
sample solution is coloured, which might be confusing for unexperienced users on the titration final point observation. Finally, by CE, it is possible to perform a refined quality control of the individual FFA in food, because the rate of oxidation process increases together with the degree unsaturation of FFA, which works as a peroxide decomposition reaction catalyst (Paradiso et al., 2010). Thus, CE can be considered a very useful technique for FFA determination in edible vegetable oils. Acknowledgements The authors wish to thank the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq – 475055/ 2011-0 and 301689/2011-3), Fundação de Amparo à Pesquisa do Estado de Minas Gerais of Brazil (FAPEMIG-CEX APQ 02420-11 and CEX-PPM 00205-11) and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) for fellowships and financial support.
References The LOD values (μM) were 7.91 for C18:1 9c, 6.96 for C16:0 and 9.06 for C18:2 9c, 12c, and the LOQ values (μM) were 26.35, 23.21 and 30.19 for C18:1 9c, C16:0 and C18:2 9c, 12c, respectively. Comparison between the CE and official titration methods
Phytochem. Anal. 2014, 25, 241–246
Copyright © 2014 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/pca
245
In order to check the reliability of CE method for the FFA analysis, the EVOO, VOO and SO samples were analysed using authentic duplicates and the results were compared with the official volumetric titration method of the European Commission Regulation (1991). Table 4 shows the statistical results, that is, of a non-parametric two-related-samples test (Wilcoxon test) because the normality assumption of sample difference did not apply for CE and titration. According to the results obtained by the Wilcoxon test, no evidence of significant differences between the two methodologies was observed within the 95% confidence interval because the p-value obtained was equal to 0.14 (p-value > 0.05). Moreover, the values obtained were in accordance with the vegetable oil acidity classification. Figure 2 shows the electropherograms obtained from the analyses of EVOO, VOO and SO samples. The current CE method compared with the official titration volumetric method presents advantages such as: making it possible to observe the individual amount of FFA, which might be interest for research into the effect of the FFA profile in the oxidative process in food; the use of a low amount of sample and organic solvents; reduced toxicity, because the organic solvents used by CE present lower toxicity; analysis error by CE can be decreased, because, in general, the
Aparicio R, Baeten V. 1998. Fats and oils authentication by FT-Raman. OclOl Corps Gras Li 5(4): 293–295. Balesteros MR, Tavares MFM, Ribeiro SJL, Polachini FC, Messaddeq Y, de Oliveira MAL. 2007. Determination of olive oil acidity by CE. Electrophoresis 28(20): 3731–3736. Barclay KD, Mckersie BD. 1994. Peroxidation reactions in plant membranes - effects of free fatty-acids. Lipids 29(12): 877–882. Barra PMD, Barra MM, Azevedo MS, Fett R, Micke GA, Costa ACO, de Oliveira MAL. 2012. A rapid method for monitoring total trans fatty acids (TTFA) during industrial manufacturing of Brazilian spreadable processed cheese by capillary zone electrophoresis. Food Control 23(2): 456–461. Barra PMD, Castro RDC, de Oliveira PL, Aued-Pimentel S, da Silva SA, de Oliveira MAL. 2013. An alternative method for rapid quantitative analysis of majority cis-trans fatty acids by CZE. Food Res Int 52(1): 33–41. Castro RDC, Sobrinho FS, da Gama MAS, Barra PMD, Antoniassi R, de Oliveira MAL. 2013. Analysis of the main conjugated linoleic acid (CLA) precursors (C18: 2 n-6 and C18: 3 n-3) in Brachiaria ruzizienses by capillary zone electrophoresis. Cent Eur J Chem 11(8): 1286–1296. Cayuela JA, Garcia JM, Caliani N. 2009. NIR prediction of fruit moisture, free acidity and oil content in intact olives. Grasas Aceites 60(2): 194–202. Danzer K, Currie LA. 1998. Guidelines for calibration in analytical chemistry Part 1: Fundamentals and single component calibration. Pure Appl Chem 70(40): 993–1014. De Castro PM, Barra MM, Ribeiro MCC, Aued-Pimentel S, Da Silva SA, De Oliveira MAL. 2010. Total trans fatty acid analysis in spreadable cheese by capillary zone electrophoresis. J Agr Food Chem 58(3): 1403–1409. De Oliveira AF, Barcelos MdFP, Gonçalves ED, Neto JV. 2008. Azeite de oliva: conceitos, classificação, usos e benefícios para a saúde humana. Circular Técnica, September 2008, ed. Belo Horizonte, MG. Empresa de Pesquisa Agropecuária de Minas Gerais; 1–5.
R. T. Sato et al. European Commission Regulation. 1991. The Characteristics of Olive Oil and Olive-residue Oil and on the Relevant Methods of Analysis. Commission Regulation (EEC) No 2568/91: Official Journal of the European Communities. Greenberg SM, Frazer AC. 1953. Some factors affecting the growth and development of rats fed rancid fat. J Nutr 50(4): 421–440. ICH Expert Working Group. 2005. Validation of a analytical procedures: Text and methodology Q2 (R1). ICH Harmonised Tripartite Guideline. Published in the Federal Register, 19 May 1997, Vol. 62, No. 96; 27463–27467. Kanya TCS, Rao LJ, Sastry MCS. 2007. Characterization of wax esters, free fatty alcohols and free fatty acids of crude wax from sunflower seed oil refineries. Food Chem 101(4): 1552–1557. Lai YW, Kemsley EK, Wilson RH. 1995. Quantitative-analysis of potential adulterants of extra virgin olive oil using infrared-spectroscopy. Food Chem 53(1): 95–98. Lercker G, Rodriguez-Estrada MT. 2000. Chromatographic analysis of unsaponifiable compounds of olive oils and fat-containing foods. J Chromatogr A 881(1–2): 105–129. Makahleh A, Saad B. 2011. Flow injection determination of free fatty acids in vegetable oils using capacitively coupled contactless conductivity detection. Anal Chim Acta 694(1–2): 90–94.
Mariotti E, Mascini M. 2001. Determination of extra virgin olive oil acidity by FIA-titration. Food Chem 73(2): 235–238. Nouros PG, Georgiou CA, Polissiou MG. 1997. Automated flow injection spectrophotometric non-aqueous titrimetric determination of the free fatty acid content of olive oil. Anal Chim Acta 351(1–3): 291–297. Osawa CC, Goncalves LAG, Ragazzi S. 2006. Potentiometric titration applied to free fatty acid determination of edible oils and fats. Quim Nova 29(3): 593–599. Paradiso VM, Gomes T, Nasti R, Caponio F, Summo C. 2010. Effects of free fatty acids on the oxidative processes in purified olive oil. Food Res Int 43(5): 1389–1394. Porto BLS, de Souza MVN, de Oliveira MAL. 2011. Analysis of omega 3 fatty acid in natural and enriched chicken eggs by capillary zone electrophoresis. Anal Sci 27(8): 872–872. Takamura K, Fuse T, Arai K, Kusu F. 1999. A review of a new voltammetric method for determining acids. J Electroanal Chem 468(1): 53–63. Zhi ZL, Rios A, Valcarcel M. 1996. An automated flow-reversal injection liquid–liquid extraction approach to the direct determination of total free fatty acids in olive oils. Anal Chim Acta 318(2): 187–194.
246 wileyonlinelibrary.com/journal/pca
Copyright © 2014 John Wiley & Sons, Ltd.
Phytochem. Anal. 2014, 25, 241–246
Revised: 20 November 2013,
Accepted: 26 November 2013
Published online in Wiley Online Library: 23 January 2014
(wileyonlinelibrary.com) DOI 10.1002/pca.2499
Rapid Separation of Free Fatty Acids in Vegetable Oils by Capillary Zone Electrophoresis Renata Takabayashi Sato,a Renata de Jesus Coelho Castro,a Patrícia Mendonça de Castro Barraa and Marcone Augusto Leal de Oliveiraa* ABSTRACT: Introduction – Olive oil is a very important product to human health since it inhibits formation of free radicals, tumour growth, lesions and inflammatory substances. High concentrations of free fatty acids in olive oils results in lipid deterioration due to oxidative or hydrolytic rancidity. Objective – To optimise an alternative capillary zone electrophoresis methodology, under ultraviolet indirect detection and to determine free fatty acids in edible vegetable oils without derivatisation steps in sample preparation. Methods – The condition used consisted of 15 mM NaH2PO4–Na2HPO4 at pH ~6.86, 4.0 mM of sodium dodecybenzenesulphonate, 8.3 mM of Brij 35®, 45% v/v of acetonitrile and 2.1% of 1-octanol, injection at 12.0 mbar of pressure for 4 s, +19 kV of applied voltage and indirect detection at 224 nm, within an analysis time of 11 min. Results – The capillary zone electrophoresis method was successfully applied to determination of free fatty acids in extra virgin olive oil, virgin olive oil and soybean oil samples. The comparison with the official volumetric titration method showed no significant difference within the 95% confidence interval. Conclusion – The main advantage to the proposed method is the possibility to observe the individual amount of the free fatty acids, which would be useful for researchers interested in studying the effect of the free fatty acids profile on oxidative process in food. Copyright © 2014 John Wiley & Sons, Ltd. Keywords: Capillary zone electrophoresis; free fatty acids; vegetable oil
Introduction
Phytochem. Anal. 2014, 25, 241–246
* Correspondence to: M. A. L. de Oliveira, Departamento de Química, Instituto de Ciências Exatas, Universidade Federal de Juiz de Fora, Cidade Universitária CEP 36036-330, Juiz de Fora, MG, Brazil. E-mail: [email protected] a
Departamento de Química, Instituto de Ciências Exatas, Universidade Federal de Juiz de Fora, Cidade Universitária CEP 36036-330, Juiz de Fora, MG, Brazil
Copyright © 2014 John Wiley & Sons, Ltd.
241
Olive oil is a very important product to human health because it inhibits formation of free radicals, tumour growth, lesions and inflammatory substances, improves the lipidic profile and prevents ageing of cells (de Oliveira et al., 2008). The best quality olive oil is the extra virgin, which is prepared from the first cold pressing of the olives and has a higher commercial price compared with the other oils. In many countries, rigid rules have been created for the consumer’s protection, taking into account different analytical possibilities to identify the authenticity and quality of the olive oil. One of the parameters evaluated for olive oil quality is the free acidity, which is expressed through the oleic acid (C18:1 9c) percentage (Table 1) (European Commission Regulation, 1991). The hydrolytic rancidity is produced from the hydrolysis of the ester bond by lipase or by moisture, resulting in the production of free fatty acids (FFA). The FFA formed are responsible for the taste and unpleasant odour, especially in fats such as butter, which has a big amount of low mass molecular fatty acid (FA). However, in fats containing non-volatile FA, such as in edible vegetable oils, the characteristic taste and odour do not appear together with the deterioration (Osawa et al., 2006). High concentration of FFA in edible oils results in lipid deterioration, which constitutes one of the most important problems in the olive oil industry. Studies performed with rats have shown that the presence of rancid food in diet might contribute to partial inactivation or destruction of vitamins A, D, E and K, carotene, tocopherols,
ascorbic acid and essentials fatty acids; in addition, it changes the digestive system operation (Greenberg and Frazer, 1953). On the other hand, the FFA accumulation in plants, for example, might indicate that they were exposed to environmental stress such as freezing or desiccation (Barclay and Mckersie, 1994). Moreover, Barclay and Mckersie’s (1994) work has shown that the determination of individual FFA is also important because the free palmitic acid may reduce aldehyde production, while free linolenic acid increases the aldehyde and decreases the fluorescent product simultaneously. Within this context, the development and optimisation of fast and reliable alternative analytical methods for oil acidity determination is very necessary. It is important to highlight that such methods can be useful to help in quality control or in the monitoring of acidity in stored olives, the extracted olive oil or even to attest the quality of the product available for consumption on the market shelf. The classic method for olive oil acidity determination is alkaline volumetric titration (AVT) (European Commission Regulation,
R. T. Sato et al. Table 1. Vegetable oil classification according to acidity Product Extra virgin olive oil Virgin olive oil Common virgin olive oil Refined olive oil Olive oil Lampante olive oil
Acidity, expressed in oleic acid (g/100 g) 0.8 2.0 3.3 0.3 1.0 >3.3
1991; Mariotti and Mascini, 2001). However, other analytical approaches have been used, such as flow injection analysis in automated systems (Nouros et al., 1997; Mariotti and Mascini, 2001), flow-reversal injection liquid–liquid extraction (Zhi et al., 1996), voltammetric reduction of quinones (Takamura et al., 1999), infrared and near infrared spectroscopy (Lai et al., 1995; Cayuela et al., 2009), Raman spectroscopy (Aparicio and Baeten, 1998), flow injection analysis (FIA) (Makahleh and Saad, 2011) and gas chromatography (Lercker and Rodriguz-Estrada, 2000; Kanya et al., 2007). Since the 1990s, capillary electrophoresis (CE), a separation technique based on differentiatial migration of neutral, ionic and ionisable compounds by application of an electrical field in a fused silica capillary tube filled with a convenient electrolyte solution has been successfully applied as an alternative method to FA analysis in different matrixes (Balesteros et al., 2007; De Castro et al., 2010; Porto et al., 2011; Barra et al., 2012, 2013; Castro et al., 2013). This present study takes into account work performed by Balesteros et al. (2007), which was the first literature account of olive oil acidity analysis by capillary zone electrophoresis (CZE). However, at the time, the methodology proposed was not so attractive for users with little experience of the CE technique, due to the necessity of removing the external polyimide coating in the ends of the capillary in order to avoid adsorption problems. This phenomenon was observed after detailed investigation in which the coating became partially dissolved in the electrolyte and the polymer gradually adhered to the capillary internal wall, disturbing the separation profile significantly and consequently decreasing the lifetime of the capillary. Furthermore, the method was applied only to extra virgin olive oil (EVOO) acidity determination. Within this context, the main goal of the present work was to demonstrate the potential of CE for FFA analysis in vegetable oils, such as EVOO, virgin olive oil (VOO) and soybean oil (SO) samples. In addition, because the method makes it possible to determine individual FFA, the oleic acid may be used as a marker to indicate possible fraudulant EVOO and, consequently, would be useful in product quality control. The proposed method therefore can be proposed as an attractive alternative to the official method by AVT (European Commission Regulation, 1991), which uses phenolphthalein as indicator, in terms of good sampling rate, economy of solvents and sample and reducing human intervention.
Materials and methods Chemicals and materials
242
All the reagents were of analytical grade and the water was purified by deionisation (Milli-Q system; Millipore, Bedford, MA, USA). The solvents methanol (MeOH), ethanol (EtOH) (Vetec, Rio de Janeiro, Brazil), acetonitrile (ACN) and 1-octanol (Merck,
wileyonlinelibrary.com/journal/pca
Rio de Janeiro, Brazil) of chromatographic grade were purchased. Polyoxyethylene 23 lauryl ether (Brij 35®) and sodium dodecylbenzenesulphonate (SDBS) were obtained from SigmaAldrich (St Louis, MO, USA), and sodium hydroxide (NaOH) was obtained from Synth (São Paulo, Brazil). Tridecanoic (C13:0), palmitic (C16:0), oleic (C18:1 9c) and linoleic (C18:2 9c, 12c) FA standards were purchased from Sigma-Aldrich. Individual FA stock solutions at a concentration of 20.0 mM were prepared by dissolving appropriate amounts of the above-mentioned standards in methanol and they were then stored in a freezer until analysis. A mixture of all the standards was prepared at concentration of 0.5 mM by the appropriate dilutions in MeOH. The aqueous Brij 35® stock solution was prepared by weighing and dissolving an amount corresponding to 50.0 mM in a 100.0 mL volumetric flask. A mass of NaOH corresponding to 1.0 M was weighed and dissolved in a 100.0 mL volumetric flask and the volume was filled with deionised water. An aqueous SDBS stock solution was prepared by weighing and dissolving a mass corresponding to 100.0 mM in a 100.0 mL volumetric flask. The aqueous buffer (pH ~6.86) stock solutions at concentrations of 100.0 mM were prepared from a mass of sodium phosphate monobasic (NaH2PO4) corresponding to 50.0 mM and 50.0 mM of sodium phosphate dibasic (Na2HPO4), which was weighed and dissolved in a 250.0 mL volumetric flask. Phosphate buffers and the Brij 35® stock solutions were kept in a freezer to prevent mould formation. The fresh working electrolyte solution was prepared by the appropriate dilutions of stocks and the incorporation of solvents. Instrumentation for the capillary electrophoresis system The experiments were conducted using a CE system (HP3d CE, Agilent Technologies, Palo Alto, CA, USA) equipped with a diodearray detector (DAD), with ultraviolet (UV) indirect detection, at 224 nm, a temperature control device (set at 25 °C), data acquisition and treatment software (HP ChemStation, Rev. A.06.01). Samples were hydrodynamically injected (12.0 mbar for 4 s), and the electrophoretic system was operated under normal polarity and constant voltage (+19 kV); manual integration was performed using peak baselines. For all the experiments, a fused-silica capillary tube with fluoropolymer (TSH) external coating was used (Polymicro Technologies, Phoenix, AZ, USA): 48.5 cm long (40 cm effective length), 75 μm internal diameter (ID) and 375 μm outside diameter (OD). In the present case, the TSH capillary avoids irreversible deleterious adsorption into the internal capillary wall, which causes intense problems in the separation performance, as demonstrated by Balesteros et al. (2007). Sample preparation Capillary Electrophoresis. The EVOO, VOO and SO samples were purchased in local market. A key point in the sample preparation step is the weighed mass: the weighed mass of each sample takes into account the separation resolution profile presented. Thus, the higher the expected acidity, the smaller should be the mass to carry the weight: EVOO mass weight was 0.75 g, VOO mass weight was 0.50 g and SO mass weighted was 1.50 g. The weighed samples were transferred to 5.0 mL volumetric flask and 0.50 mM of C13:0 was added, as internal standard (IS), and the volume was completed with ethanol at 60 °C. The
Copyright © 2014 John Wiley & Sons, Ltd.
Phytochem. Anal. 2014, 25, 241–246
Analysis of Free Fatty Acids in Vegetable Oil By Cze volumetric flask was shaken manually and after cooling the volume was completed, again, with ethanol at room temperature. It was shaken by vortex for 2 min and maintained for 2 min at rest. An aliquot of the ethanolic phase (top) was injected into the CE equipment. The analyses were performed in authentic duplicates. It is important to mention that dissolution sample tests using other solvents, such as methanol and isopropanol were performed. However, the results were not satisfactory, because with these solvents higher signal-to-noise ratios were observed in the electropherograms profiles when compared with the use of ethanol. Titration. The analyses were performed using the official cold volumetric titration method according to European Commission Regulation (1991). The samples were titrated with NaOH 0.1 M, phenolphthalein as an indicator and toluene:ethanol mixture (1:1 v/v) as the extractor reagent. The analyses were performed in authentic duplicates. The sample preparation procedure for CE method and the titration are illustrated in Fig. 1.
Capillary electrophoresis procedures Before use, new capillaries were conditioned by pressure flushing with 1.0 mM NaOH (40 min), deionised water (15 min) and electrolyte solution (15 min). The capillaries were regenerated between runs by washing with 1.0 M NaOH (2 min), deionised water (2 min) and fresh electrolyte solution (2 min). This conditioning procedure was critical to ensure peak area and migration time repeatability and to prevent solute adsorption to the capillary wall.
Statistical analysis The statistical tests such as homoscedasticity and normality were performed using IBM®SPSS® Statistics 14.0 for Windows software. The lack of fit analysis was performed in Microsoft Office® Excel software.
Results and discussion Background electrolyte Traditionally, the analysis of FFA by CZE takes into account the background electrolyte (BGE) added to organic solvent such as ACN in order to avoid micelle formation among the FFA; pH higher than 7.0 in order to promote the carboxyl group dissociation (because FFA has pKa ≈ 5.0, they are analysed in anionic form) under counter electroosmotic flow (EOF), catodic EOF and the use of chromophoric agent (SDBS) to promote indirect detection, because saturated FFA present low molar absorptivity in the UV range. Thus, the BGE consisted of 15 mM of NaH2PO4– Na2HPO4 at pH ~6.86, 4.0 mM SDBS, 8.3 mM Brij 35®, 45% v/v ACN and 2.1% 1-octanol. Thus, the output order for each FFA takes into account the charge and size relationship. In the other words, the FFA that has a low charge and size relationship, resists the EOF less and reaches the detection window first. Response factor calculation The proposal of the FFA quantification in edible vegetable oil samples was based on a statistical study including the response factor (Rf) calculation by using C13:0 as the internal standard (IS). In order to calculate Rf, a random experiment in authentic replicates using FFA standards of C16:0, C18:1 9c and C18:2 9c, 12c in concentrations of 0.2, 0.6, 1.0, 1.4, 1.8 and 2.2 mM and IS in fixed concentration of 0.5 mM, was performed by means of the leastsquares method (LSM). It is important to note that the correct practice to evaluate the fit of the model include removal of points in each level if necessary. Table 2 shows the values used in the regression model. Thus, after careful statistical evaluation (α = 0.05), that is, the assumptions estimation such as homoscedasticity test (Levene: different numbers of replicates into the same level or Cochran: same numbers of replicates into the same level), residue normality test (Shapiro–Wilk) and lack of fit test through a priori test hypothesis (ANOVA), recommended by the IUPAC (Danzer and Currie, 1998), described in equation (1), where mi is the number of measurements, p represents the calibration points, and m is the product of p and mi, the regression model was implemented. p X
F calc ¼
S2y;x S2y
¼
mi ðy i y^i Þ2 =ðp 2Þ
i¼1
p X mi X
y ij y i
2
(1) =ðm pÞ
i¼1 j¼h
In the present case, the regression model diagnosis was adequate with no lack of fit because the value of Fcalc is lower than Fcrit for all FA in the 95% confidence interval for the concentration range studied and the Rf values found were 0.526, 0.622 and 0.604 for C18:1 9c, C16:0 and C18:2 9c, 12c, respectively, as summarised in Table 3. Thus, the quantification procedure involving the calculation of Rf is described by equation (2) AFFA AC13:0 ¼ Rf ½FFA ½C13 : 0
Phytochem. Anal. 2014, 25, 241–246
where AFFA is the area for each FFA, AC13:0 is the IS area, [FFA] is the concentration (in mM) for each FA and [C13:0] is the IS concentration fixed in 0.5 mM.
Copyright © 2014 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/pca
243
Figure 1. Schemes of sample preparations for the analysis of free fatty acids by capillary electrophoresis (CE) and volumetric titration.
(2)
R. T. Sato et al. Table 2. Values used for regression model implementation FFA
Signal first replicate
[FFA]/[IS]
C18:1 9c
1.2 2.0 2.8 3.6 4.4 0.4 1.2 2.0 2.8 0.4 2.0 2.8 3.6 4.4
C16:0
C18:2 9c, 12c
Signal second replicate
78.7 96.1 208.2 194.2 292.8 36.9 111.5 – 208.5 33.4 134.6 243.8 262.2 367.6
Signal thir replicate – – – – – – – – – 28.2 160.2 211.4 344.1 472.4
70.7 90.3 147.6 181.5 303.6 31.8 129.4 179.9 217.9 28.3 127.1 220.7 317.0 406.8
Table 3. Response factor calculation FFA
Slope
C18:1 9c C16:0 C18:2 9c, 12c
0.526 (±0.016) 0.622 (±0.024) 0.604 (±0.012)
Intercept
r
Fcalc
Fcrit
-0.143 (±0.048) -0.013 (±0.048) -0.078 (±0.037)
0.996 0.994 0.997
4.42 1.80 2.86
5.41a 6.59b 3.71c
a
Ftab(n1 = 3, n2 = 5). Ftab (n1 = 3, n2 = 4). c Ftab (n1 = 3, n2 = 10). Cochran test: Ccalc (C18:1 9c) = 0.32 , Ctab (C18:1 9c) = 0.93; Ccalc (C18:2 9c, 12c) = 0.63, Ctab (C18:2 9c, 12c) = 0.79; Levene test (p-value): (C16:0): 0.06; Shapiro–Wilk test (p-values): C18:1 9c = 0.636; C16:0 = 0.378; C18:2 9c, 12c = 0.041. b
As the regression model diagnosis was satisfactory, the slope can be used as a response factor in equation (2), and as long as the internal standard C13:0 at 0.5 mM is used, the concentration of FFA remains hidden. Thus, the acidity percentage in the sample was carried out through equation (3), obtained after rearranging equation (2). It is important to note that the CZE method allows the determination of individual FFA, but the % acidity is expressed in oleic acid percentage, because this is the majority FA in olive oils. However, to calculate the % acidity value, we must take into account the sum of all the FFA present in the sample areas. %acidity ¼
MMC18:0 ½C18 : 0V Rf AC13:0 m
X
Axi
100
(3)
Table 4. Comparison between capillary electrophoresis and volumetric titration methods Sample EVOO VOO SO
CE/%oleic
244
0.35 0.40 1.68 1.65 0.07 0.07
wileyonlinelibrary.com/journal/pca
acid
Titration/%oleic 0.36 0.35 1.20 1.20 0.07 0.07
acid
where Axi is the area for each FFA in samples (C18:1 9c, C16:0, C18:2 9c, 12c), AC13:0 is the IS area, [C13:0] is the IS fixed concentration of 0.5 mM, V is the volume in litres, m is the sample mass in milligrams, Rf is the response factor (fitted model slope) and MMC18:1 9c is the molecular mass for C18:1 9c (282.5 g/mol). This approach is very interesting, because Rf is calculated just once for each FFA in preliminary experiments (under controlled operational conditions) with standards. Thus, after rigorous statistical evaluation of the regression models through the homoscedasticity, residues and linearity test evaluation through the lack of fit test (ANOVA), which did not present significant evidence for the considered interval (in the present case all values obtained in statistical tests were considered acceptable within the 95% and 96% confidence intervals), the obtained slopes can be used as Rf for the sample quantification. On other words, if some analyses were performed using the same IS, in the same concentration of the preliminary experiment and using the same operational conditions (capillary dimensions, wavelength, cartridge temperature, etc.) and there were no violations (or meaningful evidence) in the statistic tests, the calculated Rf for each FFA can be used for different samples without the necessity to perform experiments to obtain the Rf every time.
Limit of detection and limit of quantification Calculation of the limits of detection (LOD) and quantification (LOQ) can be undertaken in three different ways: (i) standard
Copyright © 2014 John Wiley & Sons, Ltd.
Phytochem. Anal. 2014, 25, 241–246
Analysis of Free Fatty Acids in Vegetable Oil By Cze
Figure 2. Sample electropherograms. Peaks: (1) C18:1 9c; (2) C16:0; (3) C18:2 9c, 12c; (IS) C13:0. Electrolyte composition: 15.0 mM of sodium phosphate buffer (pH ~6.8); 8.3 mM of Brij 35®; 4.0 mM of SDBS; 2.1% of 1-octanol and 45% of ACN. Operational conditions: 4 s injection at 12.0 mbar pressure, +19 kV applied voltage, indirect detection at 224 nm and TSH capillary 40 cm effective length.
deviation of the response and the slope, (ii) signal-to-noise approach and (iii) visual evaluation. In the present case, the signal-to-noise approach was used. This approach can be applied only to analytical procedures that exhibit baseline noise, which the signal-noise relation through the standard deviation calculation of the baseline (noise) and the height of the peak (signal) of the analytes (ICH Expert Working Group, 2005). The LOD and LOQ were calculated within the concentration corresponding to a signal-noise relationship equal to 3 (equation (4)) and 10 (equation (5)), respectively, where Sb is the standard deviation of the baseline, CS is the concentration of analyte, Hmax is the maximum peak height and Hmin is the baseline level. LOD ¼
3 Sb C S H max H min
(4)
LOQ ¼
10 Sb C S H max H min
(5)
sample solution is coloured, which might be confusing for unexperienced users on the titration final point observation. Finally, by CE, it is possible to perform a refined quality control of the individual FFA in food, because the rate of oxidation process increases together with the degree unsaturation of FFA, which works as a peroxide decomposition reaction catalyst (Paradiso et al., 2010). Thus, CE can be considered a very useful technique for FFA determination in edible vegetable oils. Acknowledgements The authors wish to thank the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq – 475055/ 2011-0 and 301689/2011-3), Fundação de Amparo à Pesquisa do Estado de Minas Gerais of Brazil (FAPEMIG-CEX APQ 02420-11 and CEX-PPM 00205-11) and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) for fellowships and financial support.
References The LOD values (μM) were 7.91 for C18:1 9c, 6.96 for C16:0 and 9.06 for C18:2 9c, 12c, and the LOQ values (μM) were 26.35, 23.21 and 30.19 for C18:1 9c, C16:0 and C18:2 9c, 12c, respectively. Comparison between the CE and official titration methods
Phytochem. Anal. 2014, 25, 241–246
Copyright © 2014 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/pca
245
In order to check the reliability of CE method for the FFA analysis, the EVOO, VOO and SO samples were analysed using authentic duplicates and the results were compared with the official volumetric titration method of the European Commission Regulation (1991). Table 4 shows the statistical results, that is, of a non-parametric two-related-samples test (Wilcoxon test) because the normality assumption of sample difference did not apply for CE and titration. According to the results obtained by the Wilcoxon test, no evidence of significant differences between the two methodologies was observed within the 95% confidence interval because the p-value obtained was equal to 0.14 (p-value > 0.05). Moreover, the values obtained were in accordance with the vegetable oil acidity classification. Figure 2 shows the electropherograms obtained from the analyses of EVOO, VOO and SO samples. The current CE method compared with the official titration volumetric method presents advantages such as: making it possible to observe the individual amount of FFA, which might be interest for research into the effect of the FFA profile in the oxidative process in food; the use of a low amount of sample and organic solvents; reduced toxicity, because the organic solvents used by CE present lower toxicity; analysis error by CE can be decreased, because, in general, the
Aparicio R, Baeten V. 1998. Fats and oils authentication by FT-Raman. OclOl Corps Gras Li 5(4): 293–295. Balesteros MR, Tavares MFM, Ribeiro SJL, Polachini FC, Messaddeq Y, de Oliveira MAL. 2007. Determination of olive oil acidity by CE. Electrophoresis 28(20): 3731–3736. Barclay KD, Mckersie BD. 1994. Peroxidation reactions in plant membranes - effects of free fatty-acids. Lipids 29(12): 877–882. Barra PMD, Barra MM, Azevedo MS, Fett R, Micke GA, Costa ACO, de Oliveira MAL. 2012. A rapid method for monitoring total trans fatty acids (TTFA) during industrial manufacturing of Brazilian spreadable processed cheese by capillary zone electrophoresis. Food Control 23(2): 456–461. Barra PMD, Castro RDC, de Oliveira PL, Aued-Pimentel S, da Silva SA, de Oliveira MAL. 2013. An alternative method for rapid quantitative analysis of majority cis-trans fatty acids by CZE. Food Res Int 52(1): 33–41. Castro RDC, Sobrinho FS, da Gama MAS, Barra PMD, Antoniassi R, de Oliveira MAL. 2013. Analysis of the main conjugated linoleic acid (CLA) precursors (C18: 2 n-6 and C18: 3 n-3) in Brachiaria ruzizienses by capillary zone electrophoresis. Cent Eur J Chem 11(8): 1286–1296. Cayuela JA, Garcia JM, Caliani N. 2009. NIR prediction of fruit moisture, free acidity and oil content in intact olives. Grasas Aceites 60(2): 194–202. Danzer K, Currie LA. 1998. Guidelines for calibration in analytical chemistry Part 1: Fundamentals and single component calibration. Pure Appl Chem 70(40): 993–1014. De Castro PM, Barra MM, Ribeiro MCC, Aued-Pimentel S, Da Silva SA, De Oliveira MAL. 2010. Total trans fatty acid analysis in spreadable cheese by capillary zone electrophoresis. J Agr Food Chem 58(3): 1403–1409. De Oliveira AF, Barcelos MdFP, Gonçalves ED, Neto JV. 2008. Azeite de oliva: conceitos, classificação, usos e benefícios para a saúde humana. Circular Técnica, September 2008, ed. Belo Horizonte, MG. Empresa de Pesquisa Agropecuária de Minas Gerais; 1–5.
R. T. Sato et al. European Commission Regulation. 1991. The Characteristics of Olive Oil and Olive-residue Oil and on the Relevant Methods of Analysis. Commission Regulation (EEC) No 2568/91: Official Journal of the European Communities. Greenberg SM, Frazer AC. 1953. Some factors affecting the growth and development of rats fed rancid fat. J Nutr 50(4): 421–440. ICH Expert Working Group. 2005. Validation of a analytical procedures: Text and methodology Q2 (R1). ICH Harmonised Tripartite Guideline. Published in the Federal Register, 19 May 1997, Vol. 62, No. 96; 27463–27467. Kanya TCS, Rao LJ, Sastry MCS. 2007. Characterization of wax esters, free fatty alcohols and free fatty acids of crude wax from sunflower seed oil refineries. Food Chem 101(4): 1552–1557. Lai YW, Kemsley EK, Wilson RH. 1995. Quantitative-analysis of potential adulterants of extra virgin olive oil using infrared-spectroscopy. Food Chem 53(1): 95–98. Lercker G, Rodriguez-Estrada MT. 2000. Chromatographic analysis of unsaponifiable compounds of olive oils and fat-containing foods. J Chromatogr A 881(1–2): 105–129. Makahleh A, Saad B. 2011. Flow injection determination of free fatty acids in vegetable oils using capacitively coupled contactless conductivity detection. Anal Chim Acta 694(1–2): 90–94.
Mariotti E, Mascini M. 2001. Determination of extra virgin olive oil acidity by FIA-titration. Food Chem 73(2): 235–238. Nouros PG, Georgiou CA, Polissiou MG. 1997. Automated flow injection spectrophotometric non-aqueous titrimetric determination of the free fatty acid content of olive oil. Anal Chim Acta 351(1–3): 291–297. Osawa CC, Goncalves LAG, Ragazzi S. 2006. Potentiometric titration applied to free fatty acid determination of edible oils and fats. Quim Nova 29(3): 593–599. Paradiso VM, Gomes T, Nasti R, Caponio F, Summo C. 2010. Effects of free fatty acids on the oxidative processes in purified olive oil. Food Res Int 43(5): 1389–1394. Porto BLS, de Souza MVN, de Oliveira MAL. 2011. Analysis of omega 3 fatty acid in natural and enriched chicken eggs by capillary zone electrophoresis. Anal Sci 27(8): 872–872. Takamura K, Fuse T, Arai K, Kusu F. 1999. A review of a new voltammetric method for determining acids. J Electroanal Chem 468(1): 53–63. Zhi ZL, Rios A, Valcarcel M. 1996. An automated flow-reversal injection liquid–liquid extraction approach to the direct determination of total free fatty acids in olive oils. Anal Chim Acta 318(2): 187–194.
246 wileyonlinelibrary.com/journal/pca
Copyright © 2014 John Wiley & Sons, Ltd.
Phytochem. Anal. 2014, 25, 241–246
