Extraction of amaranth seed oil using subcritical butane and sunsequent protein extraction from the cake generated in the process

2.1. Materials

In this work, amaranth grains of the caudatus specie, Oscar Blanco variety were used. They were obtained from the supplier ASOVITA. These grains were grown in Bolivia, in the department of Chuquisaca, Zudáñez province, more specifically in the Sopachuy region. Its harvest season was between the months of March and April, 2022. The grains were stored for 6 months, at an average temperature of 23 °C, and an average relative humidity of 46%. For the assays, part of the grain was germinated and dried to a humidity content of 12%, which was done to study the effect of germination on the oil quality and protein content of the grains. For grinding, a YB-2500 A brand Yunbang (China) cereal mill was used, where the germinated grain was crushed for 3 minutes at 2,500 rpm, thus obtaining a “fine” granulometry flour. On the other hand, two flours of different granulometry were obtained for the ungerminated grain, the first grinding for a time of 1.5 minutes at 2,500 rpm to obtain a coarse granulometry, and the second with a grinding time of 3 minutes at 2500 rpm to obtain a fine granulometry. All reagents used were purchased from Merck Co. (Darmstadt, Germany) or Winkler Ltda. (Santiago, Chile).

2.2. Characterization of the non-germinated amaranth grain

The characterization of the non-germinated amaranth grain was made based on the standards stipulated by the Bolivian Institute for Standardization and Quality (IBNORCA). To determine the moisture content, NB 312026 was used, NB 312027 for fat, NB 312029 for protein, NB 312030 for ash, and NB 312031 for carbohydrates.

2.3. Extraction of amaranth oil from germinated and non-germinated grain

To carry out the extraction, a 5 LB bidirectional extractor was used in a Closed-Circuit brand Extractor Solutions (USA), which consists of a tank with butane content, a column for the extraction of the raw material with a diameter of 6 and a length of 12, an oil collection column of 12 in diameter and 18 in length, a dehydration column, a compressor, a condenser, and a chiller (see, Figure 1). The previously crushed raw material was introduced into the extraction column, and a paper filter was installed on a stainless-steel grid at the bottom to retain the raw material inside the extraction column. After installing the flour, the system was subjected to a pressure of 90 PSI with Nitrogen gas, to verify eventual leaks, then the extractor was evacuated. Later butane was allowed to flow through the upper part of the extraction tank, where the amaranth flour was contained, and it was left to soak for 30 min. Then, the solvent with the extract was allowed to flow to the collection column, where the solvent was evaporated and recovered into the butane tank by employing the extractor compressor and a cooling system.

The extracted oil was left to rest for two days in a closed, dark environment, and then purified using a rotary evaporator at 35 °C and a partial vacuum of -40 kPa for 2 h. Finally, the oil was stored in amber containers at room temperature.

For ungerminated amaranth, a factorial experimental design was made considering as variables the ratio of mass of solvent/mass of raw material and the granulometry. On the other hand, for germinated amaranth, 2 extraction tests were performed with fine granulometry and ratios of 1/1 and 2/1 solvent/flour. The same methodology was used for all treatments.

2.4. Measurement of the ratio and granulometry

To measure the ratio, 3.0 kg was established as the standard amount of ungerminated raw material and 2.4 kg of germinated raw material to be introduced into the extractor. These quantities were selected to fill the column where the flour had to be disposed. Then, the ratio was defined by the following equation:

To determine the solvent mass needed to satisfy the ratios required, the values of the ratio and the flour mass were simply substituted in the equation previously presented in this section, and then the solvent mass was calculated. A balanced was used to ensure that the solvent mass introduced into the extractor was the calculated one. Initially, the weight of the butane container was measured and then a mathematical subtraction of this value was done, according to the following equation:

Then, the solvent was allowed to flow and the mass of the container was controlled until the obtention of the final mass calculated.

To determine the granulometry of the flours, the NB 39012 standard was used, which is based on the sifting of the flour.

2.5. Performance and Productivity Indicators

To characterize the results of the tests, two indicators were established to express the effectiveness of the process. On the one hand, the extraction yield relates the amount of oil extracted to the oil content of the grain, as given in the following equation:

This indicator was used exclusively for the extraction of oil from ungerminated amaranth grain. The oil content in the flour is obtained from the analysis carried out on a sample according to the standard procedure, in this case, a Soxhlet extraction.

On the other hand, the extraction productivity was defined as the amount of oil extracted per unit mass of flour, expressed as a percentage, which would be given by the following equation:

This indicator was used exclusively for the extraction of germinated amaranth grain oil.

2.6. Experimental design

A factorial experimental design was carried out for the extraction of amaranth oil from the non-germinated grain. The factors considered were the granulometry and the ratio of mass of solvent/mass of raw material. The response variable was the extraction yield. For the first factor, 2 levels were considered a coarse and a fine granulometry. For the second, 3 levels were considered, ratios of 8/15, 16/15, and 24/15 solvent/raw material. This is shown in Table 1. Each test was done in duplicate, and the total number of tests was 12. The variables which remained constant were the pressure, with a value of 70 PSI and the soak time of the solvent with the raw material, which was 30 min.

2.7. Statistical analysis

The data was analyzed using the Minitab 2019 software, and a two-factor analysis of variance (ANOVA 2) was applied to find out if, statistically, the factors considered in the experimental design had an impact on the response variable, with a significance level of 95%. For the analysis of the normality of the residual values, an Anderson-Darling test was applied, and to determine the homogeneity of variances of the residuals, a Bartlett test was applied.

2.8. Characterization and profile of fatty acids in the oils

To characterize the germinated and non-germinated amaranth oil, 100-ml samples of both oils were taken. The characterization was carried out by the Centro de Investigaciones Químicas SRL in the town of Quillacollo (Cochabamba, Bolivia).

The physicochemical parameters determined were: humidity, based on the Bolivian standard NB 34010; density, based on NB 34021; iodine value, based on NB 34006; refractive index at 25 °C, based on NB 34003; peroxide index, based on NB 34008; and acid value, expressed as oleic acid. The fatty acid profile for both oils was determined using the GC-FID Method FAME AOAC 969.33.

Finally, an extraction of non-germinated amaranth oil was made by cold pressing, which was carried out by the company COMEDICA Bolivia. Subsequently, this oil was characterized according to the same parameters as those of the oils obtained with subcritical butane in order to make a comparison between the values obtained from both techniques and thus know how good the quality of the oil is as produced by the technique of extraction investigated. The cold pressing technique was chosen for comparison since it generally produces high-quality oils.

2.9. Extraction of protein from defatted ungerminated amaranth flour

In the extraction of protein, the effect of the particle size on the extraction productivity was considered, for which 2 tests were carried out, one with the defatted cake of coarse particle size (9.8 % humidity and 1.49 % fat by weight) and another with the defatted cake of fine particle size (9.9 % moisture and 0.72 % fat by weight). These defatted cakes were obtained from previous oil extractions made with butane. The values considered for the basic and acid pH to which the suspension was brought were obtained from the investigations of Salcedo-Chávez et al. (2002)Salcedo-Chávez B, Osuna-Castro JA, Guevara-Lara F, Domínguez-Domínguez J, Paredes-López O. 2002. Optimization of the Isoelectric Precipitation Method To Obtain Protein Isolates from Amaranth ( Amaranthus cruentus ) Seeds. J. Agric. Food Chem. 50, 6515–6520. https://doi.org/10.1021/jf020522t, Das et al. (2021)Das D, Mir NA, Chandla NK, Singh S. 2021. Combined effect of pH treatment and the extraction pH on the physicochemical, functional and rheological characteristics of amaranth (Amaranthus hypochondriacus) seed protein isolates. Food Chem. 353 (3), 1–12. https://doi.org/10.1016/j.foodchem.2021.129466 and Cordero de los Santos et al. (2005)Cordero de los Santos MY, Osuna-Castro JA, Borodanenko A, Paredes-López O. 2005. Physicochemical and functional characterisation of amaranth (amaranthus hypochondriacus) protein isolates obtained by isoelectric precipitation and micellisation. Food Sci. Technol. Int. 11 (4), 269–280. https://doi.org/10.1177/1082013205056491, since they obtained optimal yields and a quality product under these conditions.

The extraction technique used was in an alkaline medium followed by a precipitation of the protein at the isoelectric point. A random sample of 200 g of defatted cake was suspended in water (1:9, m/m), and subsequently, the solution was basified up to pH 9 by adding 1 N NaOH solution, stirring the solution with a FLOC-6 model paddle stirrer, Raypa brand (Spain) at an angular velocity of 200 rpm, at 25 °C, and for 1 h. The pH was controlled throughout the process using a pH meter Edge brand Hanna (Romania). The solution was then centrifuged in a Pro-Hospital LL model centrifuge Centurion brand (England) at 2,100 g, for 30 min and the supernatant was transferred to another container to finally store it in a Samsung brand model RT53K65145L refrigerator (South Korea) at a temperature of 5 °C overnight. The following day, the protein concentrate was acidified to pH 5 using 1 N hydrochloric acid. The pH was controlled with a Hanna (Romania) Edge pH meter, shaking the solution with a Raypa FLOC-6 model paddle stirrer (Spain) at an angular velocity of 200 rpm for 1 h. Then the solution was centrifuged with an acceleration of 2100 g, for 30 min. At the end of the process the supernatant was discarded and the decantant was recovered as protein concentrate. The protein concentrate was neutralized with 1 N sodium hydroxide, and finally, it was lyophilized with a BK - FD10PT model lyophilizer, brand Biobase (China) for 3 days.

To verify the net protein content, the Kjeldahl analysis was performed for the 2 extractions, specified by standard NB 312029. The factor used to convert from nitrogen to protein was 6.25.

The productivity of the protein extraction from the ungerminated grain was defined as a percentage of the protein mass fraction recovered per unit of defatted cake used to perform the extraction.

3.1. Physicochemical properties of ungerminated amaranth grain and particle size

The physicochemical properties of ungerminated amaranth are shown in Table 2. Krist (2020)Krist S. 2020. Vegetable Fats and Oils. Springer Nature Switzerland. https://doi.org/10.1016/s0037-6337(12)70310-8, in her book Vegetable Fats and Oils, makes a compilation of oil values in amaranth for the caudatus variety, which oscillate between 6 to 9 g of oil/100g of seed, values which agree with the result obtained of 7.24 g of oil/100 g of seed. Calderón-Vásquez (2017)Calderón-Vásquez S. 2017. Estudio Físico- Químico del Grano de Amaranto (Amaranthus caudatus), Variedad Óscar Blanco para su Aprovechamiento con Fines Industriales. Universidad Mayor de San Andres, La Paz. http://repositorio.umsa.bo/xmlui/handle/123456789/12117, who worked on the physicochemical characterization of a non-germinated amaranth grain with the same geographical origin as the grain used, obtained a value of 7.46 %, almost the same as that obtained in this study.

Regarding protein, the result obtained was 13.33 g of protein/100 g of seed, which is close to the lowest values given in the literature of 13 g/100 g of seed (Arias-Giraldo and López-Mejía, 2021Arias-Giraldo S, López-Mejía M. 2021. Uses, nutritional properties and sensory evaluation for amaranth, quinoa and grape and coffee by-products. Ing. Compet. 24 (1), 1–16. https://doi.org/10.25100/iyc.24i1.11000; Castel, 2010Castel MV. 2010. Estudio de las propiedades funcionales, tecnológicas y fisiológicas de las proteínas de amaranto. Universidad Nacional del Litoral, Santa Fe de la Vera Cruz. http://hdl.handle.net/11185/212; Hernandez and Herrerías, 1998Hernandez R, Herrerías G. 1998. Amaranto: Historia Y Promesa. Horiz. del tiempo. 1, 1–18.; Luis et al., 2018Luis GM, Hernández-Hernández RB, Caballero VP, Torres-López G, Adrián V, Martínez E, Pacheco LR. 2018. Current and potential uses of Amaranth (Amaranthus spp.). J. Negat. No Posit. Results. 3 (6), 423–436. https://doi.org/10.19230/jonnpr.2410). This may be due to the species and variety used. Calderón-Vásquez (2017)Calderón-Vásquez S. 2017. Estudio Físico- Químico del Grano de Amaranto (Amaranthus caudatus), Variedad Óscar Blanco para su Aprovechamiento con Fines Industriales. Universidad Mayor de San Andres, La Paz. http://repositorio.umsa.bo/xmlui/handle/123456789/12117, who used an amaranth grain of the same species and variety, reported a value of 14.30 g of protein/100 g of seed, a result close to that obtained in this work.

Finally, moisture, ash, and carbohydrates, with values of 9.02 g, 2.32 g, and 68.09 g/100 g of seed, respectively, fit correctly into the values given in the literature of 9-12 g/100 g of seed for the first, 2-4.6 g/100 g of seed for the second, and 66-71 g/100 g of seed for the third (Arias-Giraldo and López-Mejía, 2021Arias-Giraldo S, López-Mejía M. 2021. Uses, nutritional properties and sensory evaluation for amaranth, quinoa and grape and coffee by-products. Ing. Compet. 24 (1), 1–16. https://doi.org/10.25100/iyc.24i1.11000; Hernandez and Herrerías, 1998Hernandez R, Herrerías G. 1998. Amaranto: Historia Y Promesa. Horiz. del tiempo. 1, 1–18.; Luis et al., 2018Luis GM, Hernández-Hernández RB, Caballero VP, Torres-López G, Adrián V, Martínez E, Pacheco LR. 2018. Current and potential uses of Amaranth (Amaranthus spp.). J. Negat. No Posit. Results. 3 (6), 423–436. https://doi.org/10.19230/jonnpr.2410).

Regarding the granulometry Figure 2, it was determined that the flour called coarse had an average particle size of 393 μm, on the other hand, the average particle size of the fine flour resulted in 291 μm.

3.2. Extraction yield

The extraction yields obtained for non-germinated amaranth are shown in Table 1. It can be seen that the highest yields correspond to the replicates of treatment 6, in which a ratio of solvent to flour of 24/15 was used, a fine granulometry of 291 μm; yield values of 92.55 and 89.83%, respectively, were obtained. This result was expected, since increasing the contact surface of the grain allows for a greater penetration of the solvent through the pores of the amaranth, and increasing the solvent ratio allows for a thorough extraction of the oil. Therefore, the smaller the granulometry of the flour used and the higher the ratio, the higher the extraction yield obtained will be. This can be seen more clearly in Figure 3.

Even though treatment 6 resulted in a higher extraction yield, treatment 5 was considered to be the optimal treatment because fewer resources (time, electrical energy, and inputs) were used to obtain similar extraction yields, which were 92.5 and 89.5%, almost the same as treatment 6 of 92.6 and 89.3%. This small difference in extraction yield is due to the ratios used, since the first treatments was completely saturated and insufficient to dissolve all the oil contained in the amaranth. As the ratio increases, there is greater solvent entry, for a constant mass of amaranth, therefore, the amount of solvent that enters ends up exceeding the amount necessary to dissolve the oil contained in the amaranth.

The effect of ratio and granulometry on the extraction yield can be statistically corroborated with a two-factor analysis of variance (ANOVA 2) as shown in Table 1, for which a significance level of 5% was considered (P < 0.05). The values obtained for P were 0.032 for the granulometry, 0.007 for the ratio, and 0.343 for the interaction between ratio and granulometry, thus concluding that the granulometry and the ratio affected the extraction yield and that there was no interaction between the two variables.

Regarding the productivity of the germinated amaranth grain shown in Table 3, absolute values quite like those obtained with non-germinated amaranth were found.

No investigations have been found on the extraction of amaranth oil using subcritical butane. However, there are studies where other types of vegetable oils were extracted with this solvent, such as that of Han et al. (2016)Han JH, Wu QF, Xu B, Zhou SL, Ding F. 2016. Quality characteristics of soybean germ oil obtained by innovative subcritical butane experimental equipment. Qual. Assur. Saf. Crop. Foods. 8 (3), 369–377. https://doi.org/10.3920/QAS2015.0625, where he found that the yield in the extraction of soybean germ oil with subcritical butane was significantly equal to that obtained by carbon dioxide in the supercritical state or with hexane. Gu et al. (2017)Gu LB, Liu XN, Liu HM, Pang HL, Qin GY. 2017. Extraction of fenugreek (Trigonella foenum-graceum L.) seed oil using subcritical butane: Characterization and process optimization. Molec. 22 (2), 1–14. https://doi.org/10.3390/molecules22020228, studied the extraction of fenugreek seed oil using subcritical butane as a solvent and also found high extraction yields.

The extraction yields found in this research are similar to those found with other extraction techniques, such as that of Westerman et al. (2006)Westerman D, Santos RCD, Bosley JA, Rogers JS, & Al-Duri B. 2006. Extraction of Amaranth seed oil by supercritical carbon dioxide. J. Supercrit. Fluids. 37 (1), 38–52. https://doi.org/10.1016/j.supflu.2005.06.012, who studied the extraction yield of amaranth oil using carbon dioxide in a supercritical state.

3.3. Physicochemical properties of the oils obtained.

The physicochemical properties of the ungerminated and germinated amaranth oil obtained with subcritical butane and the ungerminated oil obtained by cold pressing can be observed in Table 4.

Making a quality comparison between the germinated and non-germinated grain amaranth oils obtained with subcritical butane, it is notable that the quality of the non-germinated with respect to germinated Amaranth grains is highly superior. This represented the great difference that exists between the peroxide indices of 3.78 for the first and 0.47 for the second, iodine 92.98 for the first and 106.93 for the second, and acidity 13.93 for the first and 0.52 for the second. This difference may be due to a degradation in the oil of the grain during the germination process and subsequent processes of drying and grinding.

Comparing the quality of non-germinated amaranth oil obtained with subcritical butane and that obtained by cold pressing, the following is notorious. Starting with the acid value of 0.52 for the first and 1.17 for the second and the peroxide value of 0.47 for the first and 3.04 for the second, a greater amount of free fatty acids and oxidation through cold pressing were evident. this may be due to the fact that, with cold pressing, there is a heating of the flour due to the friction and high pressures that may reach values above 80 °C. They are considerably higher than those during extraction with butane, which ranged between 10 to 15 °C. In addition, throughout the extraction process with butane, neither the flour nor the oil came into direct contact with ambient air. On the other hand, almost no water was extracted during extraction with butane, and the oil is quite dry in comparison with the oil extracted with cold pressing where significant amounts of water were extracted together with oil. Regarding the iodine value, a value of 106.93 was obtained for the first and 70.12 for the second, which indicates that a greater amount of unsaturation was obtained using butane as a solvent.

3.4. Fatty acid profile of the oils obtained.

The fatty acid profile obtained for the different oils is shown in Table 4. In the three of them, stearic, palmitic, oleic, and linoleic acids stand out, with values of 7.87, 17.32, 26.04, and 21.48%, respectively, for the extraction of germinated grain with butane; with values of 4.47, 22.51, 30.69, and 40.26%, respectively, for the extraction with butane of non-germinated grain; and values of 3.54, 19.28, 28.23 46.99%, respectively, for extraction by cold pressing. It can be seen that that there is a slightly larger fraction of unsaturated fatty acids in the saponifiable matter of the oil obtained by cold pressing than in the oil obtained with butane. This is in contrast with the iodine value, which indicates that there is a greater quantity of unsaturation obtained using butane than with cold pressing. This indicates that, proportionally, the unsaturated fatty acids represent a higher proportion of the saponifiable matter in the oil obtained by cold pressing, but in absolute terms, there is a greater amount of unsaturation obtained with butane, which would indicate that the saponifiable fraction of the oil obtained with butane is greater than that obtained by cold pressing.

3.5. Extraction of protein from ungerminated amaranth cake

Table 5 shows the results obtained for the protein extraction of ungerminated amaranth cake. The fine grain, which had a smaller particle size and a smaller amount of fat, had a productivity of 5.51%, which was higher than the value obtained for the coarse grain, which was 3.57%. This result was expected, since the reduced particle size increased the contact surface, and the penetration of the basic water solution into the pores of the flour was more effective.

It is important to mention that the product obtained after lyophilization was not pure protein, instead, it was a protein concentrate. Table 5 shows the concentrates extracted from the defatted cake, with a value of 11.50 g for the fine grain, which represented 6.38% of the cake used with a net protein content of 9.29 g, and a purity of 80.78%. For the coarse grain, the protein concentrate extracted was 8.1 g, which represented 4.49% of the cake used. It had a net protein content of 6.45 g, with 79.63% purity. A mass balance was released to know the protein content in the fine and coarse grains. For the first one, it was determined to be proximately 28.1 g and the net protein extracted was 9.29 g, so it represented 33.1% of the total. On the other hand, the protein in the coarse grain was determined to be proximately 27.9 g and the net protein extracted was 6.45 g, representing 23% of the total.

Further research is necessary to determine whether the protein concentrate extracted would be considered safe for human consumption. Nevertheless, it is not because of the sub-critical butane extraction method, given that the process of lyophilization eliminates all small butane residues, but because of the process of protein isolation.

The results obtained are similar to those of Cordero de los Santos et al. (2005)Cordero de los Santos MY, Osuna-Castro JA, Borodanenko A, Paredes-López O. 2005. Physicochemical and functional characterisation of amaranth (amaranthus hypochondriacus) protein isolates obtained by isoelectric precipitation and micellisation. Food Sci. Technol. Int. 11 (4), 269–280. https://doi.org/10.1177/1082013205056491, who obtained somewhat higher amaranth protein extraction values, but with a lower pH during the precipitation stage. As shown by Salcedo-Chávez et al. (2002)Salcedo-Chávez B, Osuna-Castro JA, Guevara-Lara F, Domínguez-Domínguez J, Paredes-López O. 2002. Optimization of the Isoelectric Precipitation Method To Obtain Protein Isolates from Amaranth ( Amaranthus cruentus ) Seeds. J. Agric. Food Chem. 50, 6515–6520. https://doi.org/10.1021/jf020522t and Cordero de los Santos et al. (2005)Cordero de los Santos MY, Osuna-Castro JA, Borodanenko A, Paredes-López O. 2005. Physicochemical and functional characterisation of amaranth (amaranthus hypochondriacus) protein isolates obtained by isoelectric precipitation and micellisation. Food Sci. Technol. Int. 11 (4), 269–280. https://doi.org/10.1177/1082013205056491, protein extraction productivity can be further increased by reducing the pH to a value of 4.