Fatty acid profile and rheological behavior of annatto seed oil (Bixa orellana), cupuassu seed fat (Theobroma grandiflorum), and their blends

2.1. Materials and blend preparation

The ASO sample was provided by Gran Oils Industry (São Paulo, Brazil) and The CSF sample was provided by Amazon Oil Industry (Ananindeua, Pará, Brazil). For blend preparation, ASO and CSF samples were previously heated at 40 ºC, in order to facilitate the homogenization process.

2.2. Determination of fatty acid composition

The fatty acid profiles of ASO, CSF, and their blends were determined by gas chromatography (GC) (CP 3380, Varian Inc., USA) equipped with a flame ionization detector (FID) and a CP-Sil 88 capillary column (60 m x 0.25 mm x 0.25 mm) (Varian Inc., USA). Fatty acid methyl esters were prepared based on the method proposed by Rodrigues et al. (2010)Rodrigues AMC, Darnet S, Silva LHM. 2010. Fatty acid profiles and tocopherol contents of buriti (Mauritia flexuosa), patawa (Oenocarpus bataua), tucuma (Astrocaryum vulgare), mari (Poraquei baparaensis) nad inaja (Maximiliana maripa) fruits. J. Brazil. Chem. Soc. 21, 2000-2004. http://dx.doi.org/10.1590/S0103-50532010001000028. and the operating conditions were helium as carrier gas at a flow rate of 0.9 mL·min^-1, FID at 250 °C, and injector split ratio of 1:100 at 245 °C. The column temperature was programmed at 80 ºC for 4 min, with a subsequent increase to 220 ºC at 4 ºC·min^-1. Individual fatty acid peaks were identified by comparison of retention times of the gas chromatography with those of known mixtures of fatty acid standards (74X Nu-check-prep, Inc., USA) run under the same operating conditions.

2.3. Iodine and saponification values

Iodine and saponification values were obtained according to methods Cd 1c-85 and Cd 3-94 (AOCS, 2004AOCS. 2004. American Oil Chemists’ Society. Official methods and recommended practices of the AOCS. 4th ed. Champaing. ), respectively, based on the fatty acid compositions.

2.4. Lipid quality indexes

The nutritional quality of ASO, CSF, and their blends was evaluated using six nutritional indicators based on fatty acid compositions. Atherogenicity (AI) and thrombogenicity (TI) indexes and the ratio of polyunsaturated to saturated fatty acids (P/S) were determined according to Ulbricht and Southgate (1991)Ulbricht TLV, Southgate DAT. 1991. Coronary heart disease: seven dietary factors. Lancet 338, 985-992. https://doi.org/10.1016/0140-6736(91)91846-M. using Equations 1, 2, and 3, respectively. The ratio of hypocholesterolemic/hypercholesterolemic fatty acids (h/H) was determined according to Santos-Silva et al. (2002)Santos-Silva J, Bessa RJB, Santos-Silva F. 2002. Effect of genotype, feeding system and slaughter weight on the quality of light lamb: II Fatty acid composition of meat. Livest. Prod. Sci. 77, 187-194. https://doi.org/10.1016/S0301-6226(02)00059-3., using Equation 4. And the nutritive value index (NVI) was determined according to Chen et al. (2016), using Equation 5.

2.5. Attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR)

The spectra of ASO, CSF, and their blends were obtained with a Fourier transform infrared spectrophotometer (Shimadzu Corporation IR Prestige 21 Cat. No. 206-73600-36- Kyoto-Japan), at the 4000 to 600 cm^-1 regions, using 32 scans with spectral resolution of 4 cm^-1. OriginPro v8.0 software was used to graphically plot the spectra obtained, with the aim of evaluating possible changes in functional groups.

2.6. Rheological measurements

The flow behaviors of ASO, CSF, and their blends were determined using a programmable rotational viscometer (DV-II+, Brookfield Engineering Laboratories, USA) equipped with a DIN-87 spindle. The viscometer was coupled to an ultra-thermostatic bath (SL 152/10, Solab, Brazil) for temperature control. Measurements were performed at temperatures of 30, 40, and 50 ºC, with shear rates ranging from 26 to 155 s^-1. Shear stress (τ) and viscosity (η) were obtained using WinGather software (version 1.1, Brookfield Engineering Laboratories, USA).

3.1. Fatty acid composition, iodine value, and saponification value

The fatty acid composition, iodine and saponification values, and nutritional quality indexes of ASO, CSF, and their blends are presented in Table 1. The results indicate that ASO is rich in PUFA (49.25%) and MUFA (31.02%), whereas CSF presents high contents of saturated (42.56%) SFA and MUFA (40.60%). The main fatty acids present in ASO were α-linolenic (49.25%), oleic (31.02%), and palmitic (12.73%). Silva et al. (2008)Silva GF, Gamarra GMC, Oliveira AL, Cabral FA. 2008. Extraction of bixin from annatto seeds using supercritical carbon dioxide. Braz. J. Chem. Eng. 25, 419-426. https://doi.org/10.1590/S0104-66322008000200019 obtained similar contents for oleic (33.9%) and palmitic (16.4%) acids using this same method of analysis. However, in their work, linoleic (34.3%) was the main PUFA in ASO.

Some differences were also observed in the fatty acid composition of ASO described by Rao et al. (2015)Rao PP, Rao GN, Jyothirmayi T, Satyanarayana A, Karuna MSL, Prasad RBN. 2015. Characterisation of seed lipids from Bixa orellana and Trachyspermum copticum. J. Am. Oil Chem. Soc. 92, 1483-1490. https://doi.org/10.1007/s11746-015-2717-1.. They obtained lower contents of oleic (17.5%) and linolenic (15.1%) acids, and greater concentrations of palmitic (26.9%) and stearic (10.8%) acids. Costa et al. (2013)Costa CK, Silva CB, Lordello ALL, Zanin SMW, Dias JFG, Miguel MD, Miguel OG. 2013. Identificação de δ tocotrienol e de ácidos graxos no óleo fixo de urucum (Bixa orellana Linné). Rev. Bras. Pl. Med. 15, 508-512. https://doi.org/10.1590/S1516-05722013000400006. suggest that the variations in the fatty acid profile of ASO can be attributed to the diversity of the species studied.

Regarding CSF, the main fatty acids present were oleic (40.60%), stearic (31.78%), α-linolenic (11.56%), and palmitic (7.51%). Previous studies confirm that oleic acid is the major compound found in CSF (41.60-43.17%), followed by stearic (31.60-33.91%), α-linolenic (11.37-11.50%), and palmitic (7.37-7.50%) acids (Bezerra et al., 2017Bezerra CV, Rodrigues AMC, Oliveira PD, Silva DA, Silva LHM. 2017. Technological properties of amazonian oils and fats and their applications in the food industry. Food Chem. 221, 1466-1473. https://doi.org/10.1016/j.foodchem.2016.11.004.; Serra et al., 2019Serra JL, Rodrigues AMC, Freitas RA, Meirelles AJA, Darnet SH, Silva LHM. 2019. Alternative sources of oils and fats from Amazonian plants: Fatty acids, methyl tocols, total carotenoids and chemical composition. Food Res. Int. 116, 12-19. https://doi.org/10.1016/j.foodres.2018.12.028.). The high content of oleic acid characterizes CSF as a good source of MUFA. Bhattacharjee et al. (2020)Bhattacharjee B, Pal PK, Chattopadhyay A, Bandyopadhyay D. 2020. Oleic acid protects against cadmium induced cardiac and hepatic tissue injury in male Wistar rats: A mechanistic study. Life Sci. 244¸ 1-18. https://doi.org/10.1016/j.lfs.2020.117324. highlighted the role of oleic acid in reducing the risk of cardiovascular diseases, decreasing inflammatory responses, and regulating serum cholesterol levels. It also has the advantage of having greater oxidative stability compared to PUFA.

The main fatty acids present in the blends were oleic (33.89-37.71%), α-linolenic (22.90-38.02%), stearic (14.27-24.28%), and palmitic (9.08-11.16%). Increasing the proportion of ASO in the blends promoted an increase in the content of PUFA and a consequent reduction in SFA. On the other hand, the samples with greater amounts of CSF also presented higher contents of MUFA, which makes the blends more interesting to human nutrition due to the beneficial effects of these fatty acids on the prevention of cardiovascular diseases, besides conferring greater stability (Bezerra et al., 2017Bezerra CV, Rodrigues AMC, Oliveira PD, Silva DA, Silva LHM. 2017. Technological properties of amazonian oils and fats and their applications in the food industry. Food Chem. 221, 1466-1473. https://doi.org/10.1016/j.foodchem.2016.11.004.).

As expected, ASO showed a higher iodine value (IV) (155.52 g I₂100 g^-1) than CSF (74.30 g I₂ 100 g^-1) because it contains a greater content of unsaturated fatty acids. Thus, the use of CSF may have contributed to increasing the oxidative stability of the blends.

The saponification value (SV) is considered a measurement of the chain length of all fatty acids present in the oil or fat (Sivakanthan et al., 2019Sivakanthan S, Jayasooriya AP, Madhujith T. 2019. Optimization of the production of structured lipid by enzymatic interesterification from coconut (Cocos nucifera) and sesame (Sesamum indicum) oils using Response Surface Methodology. LWT - Food Sci. Technol. 101, 723-730. https://doi.org/10.1016/j.lwt.2018.11.085.). The SV of ASO (193.49 mg KOH g^-1) was higher than that of CSF (190.92 mg KOH g^-1).

3.2. Nutritional quality indexes

Atherogenicity and thrombogenicity indexes lower than 1.0 and 0.5, respectively, are recommended in terms of human health (Fernandes et al., 2014Fernandes CE, Vasconcelos MAS, Ribeiro MA, Sarubbo LA, Andrade SAC, Filho ABM. 2014. Nutritional and lipid profiles in marine fish species from Brazil. Food Chem. 160, 67-71. https://doi.org/10.1016/j.foodchem.2014.03.055.). ASO and CSF showed the same AI result (0.16), which indicates that these samples have high and equivalent contents of anti-thrombogenic fatty acids (Table 1). As expected, CSF showed a TI (0.66) higher than ASO (0.12), because it presents a greater content of SFA, especially stearic acid (31.78%). In addition, ASO presented a great concentration of α-linolenic acid (49.25%), which is a PUFA that reduces platelet aggregation potential. The blends showed AI and TI lower than the recommended values, which is very desirable from a human health perspective (Wołoszyn et al., 2020Wołoszyn J, Haraf G, Okruszek A, Wereńska M, Goluch Z, Teleszko M. 2020. Fatty acid profiles and health lipid indices in the breast muscles of local Polish goose varieties. Poult. Sci. 99, 1216-1224. https://doi.org/10.1016/j.psj.2019.10.02.).

The P/S ratio is frequently used to evaluate the nutritional quality of oils and fats. A balanced intake of polyunsaturated and saturated fatty acids is beneficial for controlling serum cholesterol levels. A P/S ratio higher than 0.45 is recommended in human diets for the prevention of coronary heart disease and some chronic illnesses such as cancer. The P/S ratios of all samples were consistent with the recommended value, except for CSF (P/S = 0.40) and the most favorable result was observed for ASO (2.50). In addition, all blends showed P/S results which are beneficial to human health and increasing the proportion of ASO improved the nutritional quality of the blends.

The value of the h/H ratio (hypocholesterolemic/hypercholesterolemic - h/H) ranged from 6.25 to 7.37, with the highest value for CSF and the lowest for CSO. The mixtures presented values within this range, and as the concentration of CSF in the mixture increased, this ratio increased. This relationship responds directly to the characteristics of the studied fats. The higher the value of h/H, the greater the relationship with risk factors, contributing to the increase in cholesterol, so this is an important measurement in the characterization of fats and their mixtures. It is important to have a low h/H ratio in the diet to reduce negative prothrombotic effects caused by increased n-6 linoleic acid concentration in the diet. Despite the limitations, compared to the PUFA/MUFA, the h/H ratio may more accurately reflect the effect of the fatty acid composition on cardiovascular diseases (Chen and Liu, 2020Chen J, Liu H. 2020. Nutritional indices for assessing fatty acids: a mini-review. Int. J. Mol. Sci. 21, 5695. https://doi.org/10.3390/ijms21165695.). Values as low as 0.21 can be found for algae and close to 11.0 to 15.0 for camelina oil (Camelina sativa) (Chen and Liu, 2020Chen J, Liu H. 2020. Nutritional indices for assessing fatty acids: a mini-review. Int. J. Mol. Sci. 21, 5695. https://doi.org/10.3390/ijms21165695.).

CSF exhibited a greater nutritional value index (NVI) value (9.64) than ASO (2.98), which was due to the greater proportion of stearic (C18:0) and oleic (C18:1) acids present in CSF. Furthermore, it was also observed that the addition of CSF to the blends directly influenced the NVI, which ranged from 4.31 to 6.83.

3.3. Attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR)

The spectra obtained in the region of 4,000 to 500 cm^-1 for ASO, CSF, and their blends are presented in Figure 1. The absorption bands of highest intensity were identified in the region of 2937 to 2843 cm^-1 and can be attributed to axial deformation vibrations of C-H bonds present in methyl (CH₃) and methylene (CH₂) groups, and double bonds (=C-H).

The band that appears in the 1754 - 1724 cm^-1 region refers to axial deformation vibrations of the carbonyl group (C=O) present in triacylglycerol esters. The bands with intermediate intensity, which appear in the region of 1469 to 1440 cm^-1, originate from angular deformation vibrations of C-H bonds present in methyl and methylene groups. The absorption bands which refer to the axial deformation vibrations of C-O bonds of esters which constitute the triacylglycerols are in the 1177 - 1147 cm^-1 region, which contains part of the “fingerprint” region of the compounds. (García-González et al., 2013García-González DL, Baeten V, Pierna JAF, Tena N. 2013. Handbook of Olive Oil (2nd ed). New York: Springer US, (Chapter 10).). Finally, the peaks at 719 and 711 cm^-1 can be assigned to the benzene ring.

The results obtained are consistent with the limits reported in other studies (Bitencourt et al., 2018Bitencourt APR, Duarte JL, Oliveira AEMFM., Cruz RAS, Carvalho JCT, Gomes ATA, Ferreira IM, Ribeiro-Costa RM, Silva-Júnior JOC, Fernandes CP. 2018. Preparation of aqueous nanodispersions with annatto (Bixa orellana L.) extract using an organic solvent-free and low energy method. Food Chem. 257, 196-205. https://doi.org/10.1016/j.foodchem.2018.02.067.). In addition to the bands identified in this work, a set of absorption bands in the regions of 3338 cm^-1, 1612 cm^-1, and 1562 cm^-1 was also found. These data are related to the identification of trans-bixin (carotenoid) present in the oil (Rao et al., 2014Rao MP, Manjunath K, Bhagawati ST, Thippeswamy BS. 2014. Bixin loaded solid lipid nanoparticles for enhanced hepatoprotection - Preparation, characterisation and in vivo evaluation. Int. J. Pharm. 473, 485-492. https://doi.org/10.1016/j.ijpharm.2014.07.027.).

Regarding the spectra of the blends, it was observed that the vibration bands of the bonds involved are very close and did not undergo relevant shifts with respect to wavelength, indicating the absence of chemical interactions between oil and fat (Silva-Júnior et al., 2008Silva-Júnior AA, Scarpa MV, Pestana KC, Mercuri LP, Matos JR, Oliveira AG. 2008. Thermal analysis of biodegradable microparticles containing ciprofloxacin hydrochloride obtained by spray drying technique. Thermochim. Acta 467, 91-98. https://doi.org/10.1016/j.tca.2007.10.018.). These results suggest that combining ASO and CSF in different proportions did not affect the functional groups, which is expected since it is just a mixture and there is no reaction.

3.4. Rheological behavior

The rheological curves of shear stress versus shear rate of the pure samples and their blends are presented in Figure 2. All samples showed a linear relationship between shear stress and shear rate, i.e., they exhibited Newtonian fluid behavior (Silva et al., 2017Silva DA, Santos PH, Pena RS. 2017. Colloidal and rheological behavior of aqueous dispersions of buriti tree (Mauritia flexuosa) gum. Emir J. Food Agric. 29, 716-723. https://doi.org/10.9755/ejfa.2017.v29.i9.115.). It can be noted that, for a constant strain rate, there is a decrease in shear stress values with increasing temperature (Figure 2). This effect can be associated with the structural collapse of molecules that constitute the samples, which occurs due to the action of hydrodynamic forces generated, promoting the alignment of molecules (Alparslan and Hayta, 2002Alparslan M, Hayta M. 2002. Rheological and sensory properties of pekmez (grape molasses)/tahin (sesame paste) blends. J. Food Eng. 54, 89-93. http://doi.org/10.1016/S0260-8774(01)00197-2.).

The Newtonian fluid behavior was confirmed through the parameters of experimental data that were fitted to the Newtonian model, as can be seen in Table 2. R², χ², and RSS values confirm the excellent fit of data to the Newtonian model, which indicates that this model is able to accurately predict the rheological behavior of ASO, CSF, and their blends, in the temperature range studied.

The Newtonian fluid behavior was confirmed through the parameters of experimental data that were fitted to the Newtonian model because all samples showed high R² values (> 0.998) and low χ² (< 0.003) and RSS (< 0.028) values, as can be seen in Table 3. R², χ², and RSS values confirm the excellent fit of data to the Newtonian model, which indicates that this model is able to accurately predict the rheological behavior of ASO, CSF, and their blends, in the temperature range studied.

Previous studies have attributed the Newtonian fluid behavior, characteristic of oils and fats, to the long-chain molecules that constitute them (Santos et al., 2004Santos JCO, Santos IMG, Conceição MM, Porto SL, Trindade MFS, Souza AG, Prasad S, Fernandes VJJ, Araújo AS. 2004. Thermoanalytical, kinetic and rheological parameters of commercial edible vegetable oils. J. Therm. Anal. Calorim. 75, 419-428. https://doi.org/10.1023/b:jtan.0000027128.62480.db.). Santos et al. (2005)Santos JCO, Santos IMG, Souza AG. 2005. Effect of heating and cooling on rheological parameters of edible vegetable oils. J. Food Eng. 67, 40-405. https://doi.org/10.1016/j.jfoodeng.2004.05.007. studied the rheological behavior of traditional vegetable oils and their blends and found similar results.

As shown in Figure 3, viscosity remained constant regardless of the shear rates tested, whereas the increase in temperature caused a gradual reduction in viscosity of all samples. The effect of temperature on the viscosity of oils and fats can be attributed to the reduction of intermolecular interactions that occur due to increased thermal molecular motion (Santos et al., 2005Santos JCO, Santos IMG, Souza AG. 2005. Effect of heating and cooling on rheological parameters of edible vegetable oils. J. Food Eng. 67, 40-405. https://doi.org/10.1016/j.jfoodeng.2004.05.007.).

The viscosity of all samples was inversely proportional to the SV and directly proportional to the IV. This behavior occurs because the viscosity of vegetable oils depends on the fatty acid composition, which increases according to their chain size and decreases with unsaturation (Santos et al., 2004Santos JCO, Santos IMG, Conceição MM, Porto SL, Trindade MFS, Souza AG, Prasad S, Fernandes VJJ, Araújo AS. 2004. Thermoanalytical, kinetic and rheological parameters of commercial edible vegetable oils. J. Therm. Anal. Calorim. 75, 419-428. https://doi.org/10.1023/b:jtan.0000027128.62480.db.). As expected, ASO presented lower viscosity compared to CSF at all temperatures studied because ASO has a greater content of unsaturated fatty acids (Figure 3 B). Besides, the increase in ASO caused a reduction in the viscosity of the blends. (Figure 3 A).

From an industrial point of view, the decrease in viscosity facilitates the material flow and heat exchange during processing. The lower the viscosity of a fluid, the lower the pressure drop during flow, decreasing power costs with pumping and, consequently, energy costs (Braga et al., 2013Braga ACC, Rodrigues AMC, Silva LHM, Araújo LA. 2013. Avaliação da influência da temperatura e do tratamento enzimático no comportamento reológico do suco de abacaxi pérola (Ananas comosus L. Merr.). Rev. Bras. Frutic. 35, 226-237. https://doi.org/10.1590/S0100-29452013000100026.).

The Arrhenius equation (Equation 5) adequately represented the effect of temperature on the apparent viscosity of the samples, showing R² values higher than 0.96, for both pure samples and mixtures, as can be seen in Figure 4 and Table 4.

The activation energy (E_a) indicates the sensitivity of a material to temperature changes. ASO showed lower E_a (78.99 kJ/mol) than CSF (82.03 kJ mol^-1), which indicates that the fat viscosity is more sensitive to temperature changes (Steffe, 1996Steffe JF. 1996. Rheological Methods in Food Process Engineering. 2nd ed. Freeman Press, Michigan.). The blends exhibited higher E_a results than the pure samples, indicating that the combinations of oil and fat increased their sensitivity to temperature changes.

Regardless of the method of obtaining or using vegetable oils, the knowledge of thermophysical properties, such as viscosity, is of fundamental importance for the achievement of the stages of equipment and process design or even for product specification. In the cosmetics, food and pharmaceutical industries, knowledge of such properties is necessary for the design and development of calculations, equipment and processes involving the production and formulation of new products.

The mixture of fats had a positive effect on viscosity, allowing flexibility in their applications, in addition to providing nutritional properties. While cupuassu seed fat has been used to replace cocoa fat (Medeiros et al., 2006Medeiros ML, Ayrosa AMIB, Pitombo RNM, Lannes SCS. 2006. Sorption isotherms of cocoa and cupuassu products. J. Food Eng. 73, 402-406. ), presenting a higher percentage of saturated fatty acids, annatto fat has more advantages in terms of unsaturated fatty acids, which gives the characteristic viscosity properties to two pure fats. The behavior of the mixture both contributed to a better distribution of the fatty acid profile and allowed an important variation in the viscosity of the fats.