Food Research International 47 (2012) 253258

Contents lists available at ScienceDirect

Food Research International

j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / f o o d r e s

Chemical evaluation and thermal behavior of Brazil nut oil obtained by different extraction processes
O.V. Santos a,, N.C.F. Corra b, 1, F.A.S.M. Soares a, 2, L.A. Gioielli a, 2, C.E.F. Costa b, 1, S.C.S. Lannes a, 2
a b

Faculdade de Cincias Farmacuticas, Universidade de So Paulo, Av. Prof. Lineu Prestes, 580 Bloco 16, 05508-900, So Paulo, SP, Brasil Instituto de Tecnologia (ITEC), Universidade Federal do Par, Rua Augusto Corra, n. 1, Guam, 66075-900, Belm, PA, Brasil

a r t i c l e
Keywords: Oil Brazil nut Extraction

i n f o

a b s t r a c t
Brazil nut richness has been widely studied aiming at nding transformation models that increase its yield, keep its quality, reduce costs, and increase its productivity. The objective of this study is to evaluate the spectroscopic prole and the chemical and thermal behavior of Brazil nut oil obtained by different extraction processes. Lipid soxhlet extractions with petroleum ether and hexane, and using hydraulic pressing and supercritical carbon dioxide (CO2) were performed. The physicochemical analyses showed an acidity index with signicant differences between the samples obtained by hydraulic pressing and supercritical CO2. The peroxide index presented signicant difference for the oil obtained by hydraulic pressing compared to that obtained by the other extraction processes. The fatty acid proles showed a relevant presence of the omega series (w-3, w-6, and w-9) with signicant differences for the extraction processes evaluated. The results of the thermogravimetric and differential analyses did not indicate changes in mass loss and presented predominantly exothermic behavior. The spectroscopic proles indicate that the extraction processes do not change the spectral prole of the oils. 2011 Elsevier Ltd. All rights reserved.

1. Introduction Brazil nut is one of the most important oleaginous seeds from the Amazon region. Its nutritional richness is internationally acknowledged. Therefore, it has been extensively researched by different industrial segments and attracted heavy investments aiming at isolating its major nutritional and functional products. One of the major industrial interests is on its lipid portion focusing on investigating more protable and economical isolation and/or extraction. Such investigation is based on cost-benet ratio. The extraction techniques that are based on the maximum recovery of the lipid portion of those oleaginous seeds have fostered investigations aiming at nding the optimum processing conditions. This has promoted the development of several transformation models that increase yields, reduce costs, and improve the productivity of raw material, products, and by-products thus resulting in lower processing costs and minimizing environmental residue release while maximizing yields (Freitas, Silva, Miranda, & Zanur, 2007). The seeds of the Brazil nut tree (Bertholletia excelsea H. B. K.), also known as Brazil nut in the international market, belongs to the

Corresponding author. Tel.: + 55 11 3091 3691/3650; fax: + 55 11 3815 6386. E-mail addresses: orquideavs@usp.br (O.V. Santos), nadia@ufpa.br (N.C.F. Corra), fabimarti@usp.br (F.A.S.M. Soares), lagio@usp.br (L.A. Gioielli), emmerson@ufpa.br (C.E.F. Costa), scslan@usp.br (S.C.S. Lannes). 1 Tel.: + 55 91 3201 8055; fax: + 55 91 3201 7291. 2 Tel.: + 55 11 3091 3691/3650; fax: + 55 11 3815 6386. 0963-9969/$ see front matter 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.foodres.2011.06.038

Lecythidaceae. family. It can be found in several Latin American regions, especially in the Amazon region. It has considerable potential use and is considered one of the major Amazon oleaginous seeds (Chunhieng, Haidi, Pioch, Brohier, & Montet, 2008; Santos, Lopes, Azevedo, & Santos, 2010; Sharma, Roux, & Sathe, 2009). Its lipid content is around 60 to 70% and its fatty acid prole consists mostly of unsaturated fatty acids, especially oleic, linoleic, and linolenic acids, attracts great interest (Freitas et al., 2007; Queiroga Neto et al., 2009). The processes of extraction of the richness of Brazil nut has led to more research in the search for transformation models that increase yields, preserve quality, reduce costs, and improve productivity. Among the extraction processes, the most commonly used for several oleaginous seeds in industrial scale is cold or hot old extraction using hydraulic or mechanic press with previous heating of the nuts. Other processes in laboratory scale have been employed, such as solidliquid extraction using petroleum ether, hexane, ethanol, methanol, and chloroform among other reagents isolated or combined, and also the extraction using supercritical carbon dioxide (CO2) (Freitas et al., 2007; Rodrigues, Arajo, Azevedo, & Machado, 2005; Temelli, 2009). In addition to the lipid portion extraction processes, techniques to preserve quality and evaluate the thermal characteristics of these raw materials have been the subject of intense research. Hence, those studies focus on the stability analyses due to the progressive increase in temperature monitoring the variations in the spectroscopic proles and possible changes in the products due to the peculiarity of the extraction methods and conditions used (Silverstain & Webster, 2000; Albuquerque et al., 2003).

254

O.V. Santos et al. / Food Research International 47 (2012) 253258

The objective of this study is to compare different solidliquid processes of extraction of Brazil nut oil using petroleum ether and hexane as reagents, hydraulic pressing, and supercritical carbon dioxide (CO2) investigating the possible chemical and thermal changes that different extraction processes can cause to the extracted oils. 2. Methodology 2.1. Materials and methods

2.3.5. Refraction index The refraction index was determined according to the AOCS Cc 725 method (AOCS, 1998) using an Abbe refractometer (Arealitic Jean), and the reading was done at room temperature, around 26 C, which was later on corrected to 25 C according to following equation: R = R + T T where: 2

In the present study, 2009 Brazil nut crop was used. The samples were previously dried in a forced air circulation oven (Fabbe). 2.2. Extraction processes A FMB SIWA 15 ton hydraulic press was used for the hydraulic pressing extraction by applying pressure on the samples for an average of 5 min, followed by vacuum ltration. For the solidliquid extraction, the samples were ground in a Tecnal mill-grinder (Willie TE-650) using Soxhlet together with the petroleum ether (Synth, boiling range of 3070 C) and hexane reagents (Synth, boiling range of 6870 C). The supercritical carbon dioxide (CO2) extraction was performed in the Supercritical Extraction Unit at Federal University of Par (UFPA) using P-4574 (CO2), carbon dioxide, purity 99,9% (White Martins) under 60 bar, 60 C, and ow of 15 l/min. After the samples were ground, the average granulometry of the particles was kept at around 0.85 mm diameter, using a Granutest sieve (12 mesh) with samples' average residual moisture around 3%. 2.3. Physicochemicalproperties of the extracted oils 2.3.1. Acidity index The acidity of the oils extracted via pressing was determined according to the AOCS Cd 3d-63 method (AOCS, 1998). 2.3.2. Saponication index The saponication index was determined according to the AOCS Cd 325 method (AOCS, 1998). The capillary and kinetics correction reference constants are provided by the manufacturer, while the time value is obtained from the equipment. 2.3.3. Peroxide index The oil peroxide index was determined according to the AOCS Cd 853 AOCS acetic acidchloroform method (AOCS, 1998). 2.3.4. Viscosity A Cannon-Fenske (SCHOTT GERATE, Type 520 23) viscometer was used for the analyses of viscosity of the oils conform to ASTM 446 and to the 3105 ISO norm at 40 C. A 200 capillary tube was used to introduce the sample. The viscosity value was calculated using Eq. (1): v = ktv Where: v k t v Knematic viscosity (cSt ou mm/s); capillary reference constant (k = 0,25); time obtained along the analysis (s); kinetics correction constant for t (v = 0,0) 1

K K K T T R

0.00035 to T = 20 C 0.00036 to T = 20, 50, 60 C 0.00037 to T 80 C temperature at which R reading is obtained standard reading T

The characterization of the prole of the fatty acids extracted from fatty acid methyl esters was made according to International Standards Organization ISO 5509 (1978). The oil samples of 1.0 g were weighed in tubes by adding 10.0 mL of n-heptanol and stirred. Next, 0.50 mL of NaOH and 2.0 mol/L methanol, and they were stirred again for 20s. After the phases were separated, the supernatant was collected for gas chromatography analysis (Milinsk, Matsushita, Visentainer, Oliveira, & Souza, 2008). The analyses were carried out in a gas chromatography using a Varian 430 GC chromatograph coupled with a computer using the Galaxie Chomatography software. The chromatographic conditions were: a SP-2560 fused-silica capillary column (SUPELCO USA) length of 100 m and 0.25 m inner diameter containing poliethilenoglicol. The operation conditions were: Split injection ratio 50:1; column temperature of 4 C per minute until 240 C for 5 min with helium as a carrier gas; isobaric pressure of 37 psi, linear velocity of 20 cm/s; helium make-up gas ow of 29 nL/min; injector temperature of 250 C (Varian CP8410 Autosampler); and detector temperature of 250 C. The quantitative composition will be determined by the comparison of the peak retention time with the respective fatty acid standard. The quantitative composition was determined by the area normalization, and was expressed as mass percentage (Daz Gamboa & Gioielli, 2003) according to the AOCS Ce 162 ofcial method (AOCS, 1997). All samples were analyzed in triplicate, and the values presented correspond to the mean of those values. 2.5. Thermalcharacterization of the Brazil nut oil The thermogravimetric and differential analyses (TGDTA) were carried out using a Shimadzu- DTG-60 H thermobalance with the following parameters: air ow of 60 mL/min, temperature of 750 C, heating ramp rate of 10 C/min, alumina crucible, and mass of 5 mg 0.5. 2.6. Spectroscopic haracterization of the Brazil nut oil The absorption spectroscopy in the infrared region was obtained using a thermo electron corporation IR 100 Spectrometer with resolution of 2 cm 1 64 scans and within the frequency spectra range from 4000 to 400 cm 1. 3. Results and discussion The results of the physicochemical analyses of the Brazil nut are shown in Table 1. In this study, the acidity index results obtained were expressed as (%AGL) in oleic acid. The variance analysis indicates signicant

The capillary reference and kinetics correction constants are provided by the manufacturer, while the time value is obtained from the equipment.

O.V. Santos et al. / Food Research International 47 (2012) 253258 Table 1 Physicochemical characterization of the Brazil nut oil in different processes of extraction. Analyses Brazil nut extraction processes % AGL Oleic acid Peroxide index Saponication index Refraction index Density (g/cm3) Hydraulic pressing 0.320 0,097a 4,057 0,304b 181.53 6,30a 1.456 0,002c 0,9116 0,002a Petroleum ether 0,238 0,000c 2.93 0,230a 184,10 1,790a 1.462 0,002ab 0.9100 0,000a Hexane 0.228 0,016c 2,99 0,200a 181.55 6,63a 1.463 0,001b 0.9088 0,000a

255

Supercritical CO2 0.8131 0.005b 3.40 0,200a 187.20 2,71a 1.458 0,002ac 0.9069 0,003a

Data represent the mean standard deviation of triplicate determinations. Different letters in the same row indicate statistical signicant difference (p b 0.05).

differences among the samples obtained by extractions by pressing and by supercritical CO2. Investigating Brazil nut and macadamia nut oils obtained by hydraulic pressing, Rodrigues et al. (2005) found acidity values expressed as percent of oleic acid with a mean between 0.24 and 0.20 (%AGL) respectively, These values are similar to those found in this study for solidliquid extractions with petroleum ether and hexane as solvents with a mean between 0.238 and 0.228 (%AGL). The authors state that other expression processes, such as the use of solvents, can increase the acidity value, as well as interfere in harvesting conditions, humidity, temperature, and storage among others. Freitas et al. (2007) found acidity values of 0.55 and 0.56. Among the extraction processes evaluated, the maximum value found was 0.3200 (%AGL), on average, obtained with the extraction by hydraulic pressing. The acidity index result obtained by lipid extraction using supercritical CO2 was 0.8131 (%AGL), on average, which proved better when compared to the other extraction processes studied. This can explained due to the interaction between the extractor agent and water percentage in the sample resulting in the interactions that increase the acidity of the oil extracted. With regard to relative acidity, Turner, King, and Mathiasson (2001), studying supercritical extractions of liposoluble vitamins, state that the amount of water present in the sample can be regarded as paradoxical; it can facilitate extraction or hinder it by preventing the contact between the supercritical uid with the material to be extracted. Water can cause effects of hydrolysis and degradation on the material since the contact of water with CO2 can increase its acidity due to the fact it is absorbed by water. With regard to variations in the acidity values in the sample, Rodrigues et al. (2005) added that storage conditions such as temperature and humidity can change the quality of the oil by increasing its free acidity. For the Amazon oleaginous seeds, these factors dene the quality of their by-products due to the peculiarities characterized by the region climate which features high temperature and humidity, not to mention the storage conditions. It was found in other studies with Bacuri that the acidity index of the oil extracted by the Bling & Dyer method from the in natura pulp and our were 4.2 and 4.5 (%AGL), respectively (Hiane, Bogo, Ramos, & Ramos Filho, 2003). These results show that the extraction process evaluated for the Brazil nut change signicantly the acidity index of the oils. Such index is one of the most important parameters of quality control in the samples. In parallel with the peculiarities of the extraction processes, the values obtained are in accordance with the standards provided by National Sanitary Surveillance Agency (ANVISA) for the cold pressed and non rened oils, for which the maximum value of acidity index is 4.0 mg KOH/g (Brasil, 2005). Another parameter used to evaluate the quality of oil was the peroxide index, which accesses the initial stages of the oxidation process and is able to determine the concentrations of hydroperoxides that are capable of oxidizing the potassium iodide quantifying its concentration. This is the most frequent used Index to determine oil and fat quality. The peroxide index values found for the Brazil nut oils present a signicant difference of 5% for the oil obtained by hydraulic pressing if compared with the others.

The peroxide index average values obtained for the solidliquid extraction using petroleum ether and hexane as solvents were 2.93 and 2.99 mEq kg 1. These values are similar to those found by Carvalho, Ogliari, Barrera-Arellano, and Block (2008) in studies with soybean oil during its storage, which average was 2.1 mEq kg 1. The peroxide index average values obtained for the oil extracted by supercritical was 3.40 mEq kg 1, which is similar to those found by Queiroga Neto et al. (2009) studying Brazil nut oil (2,.92 mEq kg 1). These differences indicate variations in the preservation of raw material and the changes that the time of storage, time of analysis, and the lipid extraction process peculiarities can cause to the oxidative stability of the oil extracted. The acidity and peroxide index results obtained for the extraction processes studied are in accordance with the standards provided by National Sanitary Surveillance Agency (ANVISA) for the cold pressed and non rened oils, for which the maximum values are in the range from15 mEq kg 1 to 4.0 mg KOH/g for the peroxide and acidity indexes, respectively (BRASIL, 2005). In the oil and fat analyses, the saponication index, can be dened as the amount of alkali necessary to saponify a given quantity of oil or fat. It is expressed as the milligrams of potassium hydroxide necessary to saponify 1 g of the sample (IAL, 1985). This is a parameter of great importance for oils in hygiene, skincare, and cosmetics industries among others. The saponication index results of the oils obtained by different extraction processes did not present signicant differences (p N 0.05) and were similar to those found by Ferreira, Silveira, Lucien, and Amaral, (2006) and Queiroga Neto et al. (2009) studying Brazil nut oil, 198.58 mg KOH/g and 187.5 mg KOH/g, respectively. In studies with pecan nut oil obtained by cold pressing, the value found was 184.3 mg KOH/g, which is close to that found in the present study (ORO, 2007). According to IAL (1985), the evaluation of the refraction index of the oils and fats is peculiar to each material, and it is related to the bonding saturation levels. It is a variable that can present different patterns due to the inuence of factors such as free fatty acid content, oxidation level, and thermal treatment. The average refraction indexes of the Brazil nut oil obtained by different extraction processes were 1.4567, 1.4627, 1.4637, and 1.4580 for the oil obtained by extraction with petroleum ether, CO2, hexane, and hydraulic pressing present signicant statistical differences (p N 0.05). These results are similar to those found by Queiroga Neto et al. (2009) and Ferreira et al., (2006) studying the characterization of Brazil nut crude oil with a refraction index average of 1.466 at the temperature of 20 C and 1.4451 at 25 C. With regard to Bacuri and pecan oils, Hiane et al., (2003) and Oro (2007) found values of 1.457 and 1.469 at the temperature of 40 C. For the Tucum oil obtained by cold pressing, the value found was 1.4651 at 20 C. Those data indicate that these oils present high content of oleic acid and that the extraction processes do not change the refraction index found for these oils. The analysis of density of the oils studied showed average values of 0.9116 (g/cm 3) for the oil extracted with petroleum ether, 0.9100

256

O.V. Santos et al. / Food Research International 47 (2012) 253258 Table 3 Comparison of unsaturated fatty acid composition of Brazil nut oil and other vegetable oils. Vegetable oils Relationship between Unsaturated/saturated fatty acids 2.89 2.89 3.08 2.96 2.00 2.23 5.10 2.348 5.27 2.58 2.48 4.25 Relationship between linoleic/oleic fatty/acids 1.035 1.357 1.517 1.139 0.25 0.22 5.3 0.017 2.462 0.932 1.225 0.066

(g/cm 3) for the oil extracted by CO2, 0.9088 (g/cm 3) for the oil extracted with hexane, and 0,9069 (g/cm 3) for the oil extracted by hydraulic pressing. There were no signicant differences among those samples (p N 0.05). Those results are similar to those found by Ferreira et al., (2006) in studies with Brazil nut and Tucum oil with average values of 0.910 (g/cm 3); and 0.910 (g/cm 3), respectively. This result indicates that the extraction processes studied did not change the density of the Brazil nut oils evaluated for such extraction processes. The physicochemical results found in this study conrm the quality of the oils obtained showing small statistical variations in the analyses carried out, which did not interfere with the quality parameters demanded by the Brazilian legislation. The results of the characterization of the fatty acids prole of the Brazil nut oils obtained by different extraction processes are shown in Table 2. The analysis of the fatty acid oils prole showed a rich unsaturated fatty acid composition with an average of 73.86%; 73.71%; 75.00%, and 75.13%, for the extractions by hydraulic pressing, petroleum ether, hexane, and CO2, respectively with an unsaturation average of 74.425% compared to the saturated fatty acids average of 25.575% of the extraction processes. The latter present high commercial value for confectionery and cosmetics industries. Among the fatty acids obtained, the most relevant ones were Palmitic, Stearic, Linoleic, and Oleic acids with average values of 13.82%, 11.24%, 33.20%, and 41.40%, respectively. Investigating the prole of fatty acids of Brazil nut oil, Queiroga Neto et al. (2009); Chunhieng et al. (2008), and Ceriani, Paiva, Gonalves, Batista, and Meirelles (2008) found the highest values for the palmitic, stearic, oleic,and linoleic acids. Their results are similar to those found in this study for the extraction processes studied. Comparing the values of fatty acid prole found by Ceriani et al. (2008) in studies with Buriti, grape, and soybean oils with other oleaginous seeds, it can be veried the predominance of unsaturated fatty acids in the range between 70 and 90%. Nevertheless, there are differences in the percentage distributions in the unsaturated fatty acid proles (Table 3) among the acids studied. For the Buriti oil, the oleic acid presented the highest unsaturation percentage, which was higher than 75%. For the grape and soybean oils, the linoleic acid presented the highest unsaturation values, which were 68.6% and 55.5%, respectively, showing the variations in the distribution of the unsaturation patterns among the fatty acids. The analysis of the relationship between the composition of unsaturated and saturated fatty acids in the Brazil nut oils obtained by different extraction process showed higher unsaturation with threefold higher values for almost all oils studied regardless of the extraction process by which they were obtained. The composition values found for the oleic and linoleic fatty acids presented a close relationship between them with a small positive variation for the linoleic fatty acid. With regard to the oils and their respective extraction processes, the highest composition values found were for the oils obtained by solidliquid extraction.

Brazil nut (pressing) Brazil nut (petroleum ether) Brazil nut (hexane) (Hexano) Brazil nut (CO2) Andiroba1 Bacuri2 Chamomile3 Tucum4 Soybean5 Brazil nut6 Brazil nut7 Buriti8

Brasamazon (2002)1; Hiane et al. (2003)2; Pereira (2008)3; Ferreira et al. (2008)4; Ceriani et al. (2008)5, 6; Queiroga Neto et al. (2009)7; and Silva et al. (2009)8.

The results obtained for the relationship between the composition of unsaturated and saturated fatty acids, as well for the composition of linoleic and oleic fatty acids in this study are similar to those found by Queiroga Neto et al. (2009) and Ceriani et al. (2008) with Brazil nut oil. When compared to other Amazon oleaginous seeds such as Bacuri (Hiane et al., 2003), Tucum (Ferreira, Silveira, Lucien, Amaral, & Silveira, 2008), and Andiroba (Brasamazon, 2002), the unsaturation relationship values between linoleic and oleic acid were also similar. Other oleaginous seeds such as Buriti (Silva et al., 2009), soybean (Ceriani et al., 2008), and chamomile (Pereira, 2008) presented unsaturation values twofold higher than those of other oleaginous seeds mentioned in terms of unsaturation relationships. With regard to the fatty acid composition values found among all oleaginous studied, the soybean presented the highest average value, 2.462, and the Tucum oil presented the lowest value of 0.017 for the linoleic acid. The importance of such relationship values reinforces the potential application of those oils. According to Kelley (2001), the unsaturated oleic and linoleic fatty acids are part of the components responsible for the immune system modulation and the organic responses to inammatory processes. James, Gibson, and Cleland (2000) studying diets with sources of polyunsaturated fatty acids such as olive, corn, sunower, and canola (Brassica napus L. e Brassica rapa L.) oils among others, reported them as producers of the inammatory mediators due to their ability to convert by the action of olongases and desaturases in arachidonic, docosapentaenoic, and eicosatrienoic acids among others. These acids can produce a variety of eicosanoids, prostaglandins and leukotrienes, which are responsible for the post-in ammatory and anti-in ammatory processes. According to Silveira et al. (2005), the composition of the fatty acid proles also present anti-microbial activities adding even higher functionality to those matrices that contain these compounds.

Table 2 Characterization fatty acids of Brazil nut oil in different processes of extraction. Fatty acids prole Extraction processes Miristic acid (C14) Palmiticacid (C16) Stearic acid (C18) Oleic acid (C18:1, W-9) Linoleic acid (C18:2,W-6) Linolenic acid (C18:3,W-3) Pressed oil 0,.058 0,00a 14.24 0,02ab 11.19 0,01a 36.26 0,03a 37.53 0,03a 0.076 0,00a Petroleum ether oil 0.054 0,001a 13.47 0,00bc 11.93 0,012b 31.24 0,02b 42.42 0,03bc 0.058 0,004b Hexane oil 0.052 0,002a 12.63 0,159c 11.63 0,141c 29.76 0,408b 45.17 0,722c 0.074 0,005a CO2 oil 0.516 0,03b 14.94 0,927a 10.23 0,007d 35.,55 2,275a 40,.51 2,56ab 0.075 0,005a

Data represent the mean standard deviation of triplicate determinations. Different letters in the same row indicate statistical signicant difference (p b 0.05).

O.V. Santos et al. / Food Research International 47 (2012) 253258

257

With regard to the thermoanalytical aspects related to the oils, the results presented by the thermogravimetric curves (TG) in air atmosphere show a relative stability until the range close to 250 C. According to the TG curve, the thermal decomposition process occurred in a single step, and the mass loss varied around 98% at the temperature range from 250 to 550 C. Hence, the oil thermal stability pattern was determined from the temperature interval in which the mass remained unaltered (Fig. 1). The changes observed in the thermogravimetrical curves of the oils indicate the presence of oxidative processes and almost complete carbonization of the material. The TG curves (Fig. 2) in air ow of the oils obtained by different extraction processes presented similar percentage values of mass loss in the different temperature rates regardless of the extraction process used. However, the deviation between the curves demonstrates possible changes. Therefore, focusing on a better visualization of the effects occurred, the rst TG-derived curves (DTG) were projected, in which the graphical effects show the events in details and with higher sensitivity level related to the TG peaks and their respective mass los. The rst TG-derived curves are shown respectively in details in its DTG. In these curves, it can be observed the events with major details. The most intense are around 350 C in the oil extracted with hexane. Simultaneous events of lesser amplitutude can also be observed in the curves that refer to the oils extracted by CO2 and petroleum ether in the temperature range between 350 and 450 C. Among all oils studied, the extraction by pressing showed the lesser alterationin the DTG curves. The differential thermal analyses (DTA) show the results of the behavior of the oils when subjected to temperature changes. The DTA curves in air ow show alterations in frequent and intense exothermic peaks in the range between 300 and 350 C. In the range between 450 and 500 C, the most evident alteration is the exothermic peak presented by the oil extracted by hexane and petroleum ether. The oils extracted with hydraulic pressing and supercritical CO2 kept an exothermic behavior with successive events of lesser intensity in the ranges between 300 and 350 C, and 400 C and 500 C. The last exothermic events were detected around 500 C (Fig. 3). Those predominantly exothermic behaviors presented by the oils result from the oxidative decomposition processes that the temperature increase causes to the lipidic matrices and are probably increased due to the high unsaturation of the Brazil nut oil. The differences in the graphical representations can result from the peculiarities of the extraction processes used to obtain the oils, such as possible minimum residues in the oil and exposure to deteriorating and oxidative factors, among others. The spectroscopic behavior of the oils obtained by different extraction processes in the infra-red region is shown in Fig. 4.

Fig. 2. DTG Curves of Brazil nut oil in different processes of extraction; (___) O2, (_ _ _) Ether, (.......) Hexane, (- - ) Pressing.

The results of the spectroscopic proles show that the extraction process does not interfere in the oils behavior. The wide bands of 3008 cm 1, 2750 cm 1, and 1750 cm 1 are characteristic of the methyl, carbonyl, and methyl ester groups, respectively, and are frequent in the long chain fatty acids. A band range around 1200 cm 1 is characteristic of the functional saturated esters (COC), and a band range between 1290 and 1040 is characteristic of fatty acids. The comparison of the spectral behavior of those oils and their possible changes is shown in Fig. 4. Conclusions The extraction processes evaluated in this study presented signicant differences in the acidity and saponication indexes with more relevant changes for the extraction by supercritical CO2 and hydraulic pressing, respectively. With regard to the fatty acids proles, the results show the variations between the fatty acid contents and their respective extraction processes. In general, they present unsaturation richness due to the strong presence of essential fatty acids based on the oleic, linoleic, and linolenic fatty acid proles. The results of the thermogravimetric and differential analyses show that the extractions studied did not change mass loss due to the peculiarities of the extraction process chosen. The mass loss follows a common pattern in all samples and temperature ranges. With regard to the DTA curves, the oils exhibited an exothermic behavior, characteristic of oxidative decompositions.

Fig. 1. TG curve in air atmoshere of Brazil nut oil in different processes of extraction; (___) CO2, (_ _ _) Ether, (.......) Hexane, (- - ) Pressing.

Fig. 3. DTA Curves of Brazil nut oil in different processes of extraction; (___) CO2, (_ _ _) Ether, (.......) Hexane, (- - ) Pressing.

258

O.V. Santos et al. / Food Research International 47 (2012) 253258 leo bruto da castanha-do-brasil (bertholletia excelsa h.b.k). Alimentos e Nutrio, 17(2), 203208. Ferreira, E. S., Silveira, C. S., Lucien, V. G., Amaral, A. S., & Silveira, C. S. (2008, out/dez). Caracterizao fsico-qumica do fruto e do leo extrado de Tucum (astrocaryum vulgare mart). Alimentos e Nutrio, 19(4), 427433. Freitas, S. P., Silva, F. O., Miranda, I. C., & Zanur, M. A. (2007, ago). Extrao e fracionamento simultneo do leo de castanha-do-brasil com etanol. [Campinas]. Cincia e Tecnologia de Alimentos, 27, 1417 [supl.]. Hiane, P. A., Bogo, D., Ramos, M. I. L., & Ramos Filho, M. M. (2003, May/Aug). Carotenides pr-vitamnicos a e composio em cidos graxos do fruto e da farinha do bacuri (scheelea phalerata mart.). Cincia e Tecnologia de Alimentos, 23(2), 206209, doi:10.1590/S0101-20612003000200018 [Links]. INSTITUTO ADOLFO LUTZ IAL (1985). leo e gorduras. Mtodos Qumicos para Anlise de Alimentos (3 ed.). [So Paulo. [Links]]. Iso International Organization For Standardization (1978). Animal and vegetable fats and oils preparation of methyl esters of fatty acids. Method ISO, 5509. (pp. 0106) Geneve: ISO [197, [Links]]. James, M. J., Gibson, J. R. A., & Cleland, L. G. (2000). Dietary polyunsaturated fatty acids and inammatory mediator production. The American Journal of Clinical Nutrition, 71(1), S343S348 [Links]. Kelley, D. S. (2001). Modulation of human immune and inammatory responses by dietary fatty. Nutrition, 17(7/8), 669673. Milinsk, M. C., Matsushita, M., Visentainer, J. V., Oliveira, C. C., & Souza, N. E. (2008). Comparative analysis of esterication methods in the quantitative determination of vegetable oil fatty acid methyl esters (FAME). Journal Brazil Chemistry Society, 19(8), 14751483 [Links]. Oro, T. (2007) Composio nutricional, compostos bioativos e vida de prateleira de noz e leo prensado a frio de noz-pec [carya illinoinensis (wangenh.) c. koch] Dissertao, 105 f. - Ps-Graduao em Cincia de Alimentos, Departamento de Cincia e Tecnologia de Alimentos do Centro de Cincias Agrrias, Universidade Federal de Santa Catarina, Florianpolis. [Links] Pereira, N. P. (2008). Estudo toqumico do leo da semente de [Camomilla Recutita (L.) Rauschert] Camomila, com avaliao de propriedades fsico-qumicas, biolgicas e funcionais em emulses. 192 f. Tese, Programa de Ps- Graduao em Cincias Farmacuticas, Universidade Federal do Paran, Curitiba, 2008. Queiroga Neto, V., Bakke, O. A., Ramos, C. M. P., Bora, P. S., Letelier, J. C., & Conceio, M. M. (2009). Brazil nut (Bertholletia excelsa H.B.K) seed kernel oil: Characterization and thermal stability. Revista Brasileira de Biologia e Farmcia, 3(2) [Links]. Rodrigues, J. E., Arajo, M. E., Azevedo, F. F. M., & Machado, N. T. (2005). Phase equilibrium measurements of Brazil nut (Bertholletia excelsa) oil in supercritical carbon dioxide. Journal of Supercritical Fluids, 34, 223239. Santos, O. V., Lopes, A. S., Azevedo, G. O., & Santos, A. C. (2010). Processing of Brazil-nut ouer: Characterization, thermal and morphological analysis. [Campinas]. Cincia e Tecnologia de Alimentos, 30(Suppl. 1), 264269 [mai. 2010]. Sharma, G. M., Roux, K. H., & Sathe, S. K. (2009). A sensitive and robust competitive enzyme-linked immunosorbent assay for Brazil nut (Bertholletia excelsea L.) Detection. Journal of Agricultural and Food Chemistry, 57(2), 769776, doi: 10.1021/jf803003z. Silva, S. M., Sampaio, K. A., Taham, T., Rocco, S. A., Ceriani, R., & Meireles, A. J. A. (2009). Characterization of oil extracted from Buriti fruit (Mauritia exuosa) grown in the Brazilian Amazon region. Journal American Oil Chemistry Society, 86(7), 611616, doi:10.1007/s11746-009-1400-9 [2009]. Silveira, C. S., Pessanha, C. M., Loureno, M. C. S., Neves, I., Jr., Menezes, F. S., & Kaplan, M. C. (2005). Atividade antimicrobiana dos frutos de Syagrus racea e Mauritia vinfera. Revista Brasileira de Farmacognosia, 15 (5), 143 148, doi:10.1590/S0102695X2005000200012. Silverstain, R. M., & Webster, F. X. (2000). Identicao espectromtrica de compostos orgnicos (6th. Ed). Rio de Janeiro: LTC. Temelli, F. (2009). Perspectives on supercritical uid processing of fats and oils. Journal of Supercritical Fluids, 47, 583590, doi:10.1016/j.supu.2008.10.014. Turner, C., King, J. W., & Mathiasson, L. (2001). Supercritical uid extraction and chromatography for fat-soluble vitamin analysis. Journal of Chromatography. A, 936, 215237 [Links].

Fig. 4. Spectroscopic bands of Brazil nut oil obtained from different processes of extraction.

The spectroscopic prole results indicate that the extraction processes do not change the behavior of the oils presenting bands characteristic of long chain fatty acids and saturated esters. References
Albuquerque, M. S., Guedes, I., Alcantara, P., Jr., & Moreira, S. G. C. (2003). Infrared absorption spectra of Buriti (Mauritia exuosa L). Vibrational Spectroscopy, 33, 127131, doi:10.1016/S0924-2031(03)00098-5. American Oil Chemists Societ (1997). Ofcial methods and recommended practices of the AOCS (5th ed.). Champaign, Illinois: AOCS. American Oil Chemists Societ (1998). Ofcial methods and recommended practices of the AOCS (5th ed.). Champaign, Illinois: AOCS. Brasamazon (2002). [Retrieved in Aug 8th]. http://www.brasamazon.com.br/arquivos. doc [from] BRASIL. Resoluo RDC/ANVISA/MS n 270 (2005,de 22 setembro de). Regulamento tcnico para leos vegetais, gorduras vegetais e creme vegetal. Dirio Ocial da Repblica Federativa do Brasil. Braslia, DF, 23 set. 2005. Seo 1. [Retrieved in Aug 11th., 2010, from]. http://www.suvisa.rn.gov.br/contentproducao/aplicacao/ sesap_suvisa/arquivos/gerados/resol_270_set_2005.pdf Carvalho, S. M., Ogliari, P. J., Barrera-Arellano, D., & Block, J. M. (2008, abr./jun.). Efeito da adio de tocoferis naturais sobre a qualidade de leo de soja renado e embalado em PET durante a estocagem. Brazil Journal Food Technology, 11(2), 134143. Ceriani, R., Paiva, F. R., Gonalves, C. B., Batista, E. A. C., & Meirelles, A. J. A. (2008). Densities and viscosities of oils of nutritional value. Journal of Chemical and Engineering Data, 53(8), 18461853, doi:10.1021/je800177e. Chunhieng, T., Haidi, A., Pioch, D., Brohier, J., & Montet, D. (2008). Detailed study of Brazil nuts (Bertholletia excelsa) oil micro-compounds: phospholipids, tocopherols and sterols. Journal Brazil Chemistry Society, 19(7), 13741380, doi:10.1590/S010350532008000700021. Daz Gamboa, O. W., & Gioielli, L. A. (2003). Consistncia de lpidos estructurados a partir de aceite de pescado y grasa de palmiste. Grasas y Aceites, 54(2), 122129. Ferreira, E. S., Silveira, C. S., Lucien, V. G., & Amaral, A. S. (2006, abr./jun.). Caracterizao fsico-qumica da amndoa, torta e composio dos cidos graxos majoritrios do