The comparison of krill oil extracted through ethanol–hexane method and subcritical method

Abstract This study aimed to develop a safe method EH (ethanol–hexane) to extract two kinds of krill oil (KO) simultaneously and analyze their composition. Meanwhile, subcritical butane and subcritical butane‐dimethyl ether extraction were used to extract KO for analysis comparison. Folch method was used to extract total lipids. When the volume ratio of ethanol to hexane is 4:6, the separation effect of ethanol layer and hexane layer is best. At this condition, the EH method yielded similar amount of lipids (up to 97. 72% of total lipids) with subcritical butane extraction method (97.60%). The recovery rate of ethanol and hexane was 83.6% and 86.86%, respectively. KO in hexane layer and extracted by the subcritical butane method are abundant in astaxanthin (910 and 940 mg/kg respectively), while KO in the ethanol layer had the highest phospholipid (PL) content (47.34%), n−3 polyunsaturated fatty acids (PUFA) content (45.51%), and the lowest fluorine content (11.17 μg/g), making it a potential candidate in the nutraceutical and antioxidant industry.

Krill oil (KO), with abundant eicosapentaenoic acid (C20:5 EPA), docosahexaenoic acid (C22:6 DHA), as well as antioxidant (Lu et al., 2017), has attracted increasing attention. EPA and DHA in KO are for the synthesis of both phospholipids (PLs) and triglycerides (TGs), while in fish oil (FO) are primarily just in the form of TGs (Liu et al., 2014;Rossmeisl et al., 2012). Some studies show that plasma levels of EPA and DHA increased with the consumption of KO rather than FO (Costanzo et al., 2016;Ramprasath, Eyal, Zchut, & Jones, 2013;Ramprasath, Inbal, Sigalit, & Jones Peter, 2014). In addition, the unique and particular amphiphilic nature of PLs brings emulsification properties to KO, thus improving the bioavailability of EPA and DHA (Schuchardt et al., 2011). Moreover, compared with FO, KO has a significant amount of astaxanthin, a potent and naturally occurring antioxidant, which can not only reduce the oxidation of KO (Hussein et al., 2007), but also provides health-promoting properties such as reducing the incidence of inflammation, cancer, diabetes, immune function, and hyperlipidemia (Feng et al., 2018;Kim et al., 2016;Sun et al., 2017).
Organic solvents and supercritical carbon dioxide  are most commonly used for industrial oil extraction (including KO; Bruheim et al., 2015;Gigliotti, Davenport, Beamer, Tou, & Jaczynski, 2011;Sahena et al., 2009;Xie et al., 2017). Though the properties of components in KO separated via the SC-CO 2 method can be improved in certain cases, high capital costs for batch extraction and engineering hardware technology should be considered (Friedrich & Pryde, 1984;Gigliotti et al., 2011). These years subcritical solvents are becoming popular to extract oil since this method is easier than the supercritical method regarding its application in industries with a higher productivity (Xu, Han, Zhou, Wu, & Ding, 2016). What's more, subcritical butane extraction has safe, efficient, and environmental compatibility. The extraction is a continuous counter current process in which the solvent can be removed completely by system depressurization (Guan, Jin, Li, Huang, & Liu, 2018). Organic solvent extraction could extract KO more simply without expensive instruments (Xie et al., 2017).
Traditionally, KO is extracted via a two-step solvent extraction, using acetone and ethanol (Beaudoin & Martin, 2004). However, these two separate extraction steps are laborious and inefficient.
The process to extract KO has been improved via one-step extraction with isochoric ethanol and acetone (Gigliotti et al., 2011), using freeze-dried krill as raw material, making oil extraction more efficient. Though freeze-drying can preserve food quality of products, it is usually used in small-scale and biopharmaceutical industries because of high expenses for equipment and energy consumption (Tang, Tian, Lee, & Row, 2012). Air convective dryers with hot air are commonly used, while can result in severe damage to food quality, such as nutrient loss, bad taste, and color deterioration (Maskan, 2001). Moreover, serious environmental issues have arisen owing to the use of environmentally unfriendly solvents.
In this study, mixture of environmentally friendly solvents ethanol and hexane was applied as solvents to obtain two kinds of KO. The effects of different conditions (ethanol/hexane ratio, time, temperature, shrimp/solvent ratio) on the lipid yield were investigated. It is reported that the extrusion of krill meal prior to oil extraction could promote lipid yield when using n-hexane as the extraction solvent (Yin et al., 2015). Hence, before the experiment, Antarctic krill were extruded until its water content reduced to 60%. This method was compared with the subcritical extraction method. The quality of KO extracted via EH and subcritical method was measured: PL, fatty acids (FA), and minor components including astaxanthin, fluorine, ash, and arsenic in the extracted KO were analyzed.

| Materials
Antarctic krill were obtained from China National Fisheries Co., Ltd.
Then, the shrimp meal was obtained after the process of Microwave thawing, heating at 95°C for 5 min, and centrifugation, followed by freezing. Frozen Antarctic krill were then delivered to our laboratory and stored at −30°C until use. Prior to the experiments, the krill were crushed using a hammer crusher and then stored in a lowtemperature (−30°C) warehouse.

| Lipid extraction
When ethanol and hexane were used as extraction solvents, the following experiments were conducted: 10 g of frozen Antarctic krill was weighed, and KO was extracted using EH (ethanol/hexane = 4:6). After extraction, the filtrate was stratified, and the upper hexane layer and the lower ethanol layer were separated. KO in the upper and lower layers was obtained via rotary evaporation and then collected in a glass vessel for analysis. The optimization experiments were shown in Supporting Information Data S1.
A pilot-scale subcritical extraction unit purchased from Henan Subcritical Bio Technology Co., Ltd. (Anyang, Henan, China) was used to conduct the subcritical extraction. The experiment was performed at 30℃ for 1 hr at a pressure range of 0.3-0.8 MPa and was repeated four times with butane and butane-dimethyl ether as solvent (Xu et al., 2016).
For comparison, Antarctic krill oil was assessed via Folch method after the Antarctic krill cells were thoroughly homogenized (Folch, Lees, & Sloane Stanley, 1957).

| Solvent recovery rate
After lipid extraction through EH method, the volume of recycled ethanol and hexane was recorded, respectively, to calculate the recovery rate. That rate was determined gravimetrically based on the following formula: where a is the volume of the recycled ethanol or hexane(v) and b is the previously added ethanol or hexane(v).

| Determination of PLs in extracted KO via nuclear magnetic resonance (NMR) analysis
The PL content was determined using 31 P NMR analysis, as previously described by Li et al. (Li, 2014) with slight modifications. All NMR ex- ). The yields of PC and PE were calculated in accordance with the peak area ratios relative to that of the TPP reference.

Lipid extraction efficiency% =
The KO extracted by EH The KO extracted by Folch × 100% was used to analyze astaxanthin based on the method described by Rao, Baskaran, Sarada, and Ravishankar (2013) with slight modifications. Briefly, the extracted KO was dissolved in 3 ml of dichloromethane. After filtration with a 0.22-micron filter, 0.5 ml of the filtrate was placed into liquid vials. The astaxanthin was analyzed via HPLC. Astaxanthin was detected at 476 nm, and the measured quantity was based on the peak area of total astaxanthin.  (Yin et al., 2015).  (2001)) with a slight modification. Hundred milligrams of KO was accurately weighed and put into a beaker and then covered with 10 ml 0.1 M perchloric acid to dissolute for 2 hr. After dissolution, the sample solution was transferred into a 50-ml volumetric flask. Total ionic strength adjustment buffer (TISAB) solution (25 ml) was added to the flask and then diluted with deionized water to the required volume. The contents were shaken well, allowed to stand for 30 min, and poured into a 100-ml plastic beaker. Thereafter, the method described by Yin et al.

| Composition analysis of FA via gas chromatography (GC)
(2017) was followed to determine fluoride content.

| Determination of arsenic content in extracted KO via inductively coupled plasma mass spectrometry (ICP-MS)
For microwave digestion treatment, 0.3 g of KO was digested with 10 ml of HNO 3 . The sample was allowed to stand for approximately 10 min to eliminate the gases generated initially by the mixture and to avoid an excessive increase in pressure in the digestion process.
For ICP-MS analysis, after digestion, the solutions (approximately 11 ml) were diluted to a final volume of 100 ml, filtered with a 0.45 μl membrane into clean polyethylene flasks, and stored in a refrigerator at 4°C until use.
Thereafter, the method of López, Garcia, Morito, and Vidal (2003) was followed to determine the arsenic content.

| Determination of free fatty acids in extracted KO via the colorimetric method
Free fatty acid (FFA) content of each oil was determined in accordance with Véroniquej, Virginia, and Jamesk (2008) with slight modification. Reagents and samples were equilibrated to 20 ± 3°C; 20 mg of KO was dissolved in 3 ml of n-heptane and stirred for 1 min. One milliliter of copper reagent was then added, the mixture was stirred for 2 min, and the developed color was measured using a colorimeter with a 715-nm filter after incubation at room temperature (25°C) for 10 min. The results were compared to a standard curve plotted from samples of oleic acid.

| Statistical analysis
The oil extraction experiments were performed in triplicate (n = 3).
For each triplicate, at least three measurements were performed, and the results were shown as the mean ± standard deviation. Statistical analysis was conducted using the statistical software SPSS version 20.0. The results were statistically evaluated by one-way analysis of variance (ANOVA) using Tukey's test. Significant differences between means were assessed at p < 0.05.
And a supervised OPLS-DA method was then applied to sharp the classification effect of KO extracted through different methods (Wang et al., 2016).

| Effect of different conditions on lipid yield via the EH method
Supporting Information Figure S1 showed that ethanol and hexane destroyed the outer structure of shrimp powder, thus increasing the dissolution of oil.
As shown in Figure 1a, when the volume of ethanol was higher than hexane, the filtrate did not stratify. As the proportion of ethanol in the extraction solvent decreased, lipid extraction efficiency increased. When the volumetric ratio of ethanol/hexane was 6:4, lipid extraction efficiency increased up to 74.49%. When the volume of hexane was higher than ethanol, after standing for 30 min, the filtrate stratified into two layers. The total lipid extraction efficiency can approach 95.23% (54.77% for ethanol layer and 40.46% for hexane layer) at a volumetric ratio of ethanol/hexane 4:6.
The effect of reaction time on lipid extraction efficiency of ethanol and hexane layers was evaluated and presented in Figure 1b.  Meanwhile, the recovery rate of ethanol was 83.6%, and that of hexane was 86.86%.

| Comparison of different kinds of KO composition
Lipid yield (%), water (%), astaxanthin (%), FFA (%), and ash (%) in KO extracted through different methods are shown in Table 1 Table 1 also showed that astaxanthin was more soluble in nonpolar solvents rather than polar solvents. Figure 2 showed the distribution of free astaxanthin and astaxanthin esters in KO. The free astaxanthin accounted for a small part.
Due to the long storage time of Antarctic krill, the content of FFA shown in Figure 1 was generally high. KO extracted by the subcritical butane method contained similar FFA content to KO Phospholipid (PL) content presented as wt% of extracted KO ( Figure 3). As shown in Figure 3a, as for the EH method, PL content of KO in the ethanol layer (47.34%) was higher than that using  C20:1, C20:5 (EPA), and C22:6 (DHA) were the primary FAs in the oil, concurrent with previous findings. (Fricke, Gercken, Schreiber, & Oehlenschläger, 1984;Phleger, Nelson, Mooney, & Nichols, 2002) Essentially, n−3 PUFA (mainly EPA and DHA in KO) were particularly abundant, accounting for more than 23% of the total FA in all samples.
Additionally, FA composition, especially n−3 PUFA, exhibited considerable differences among KO samples extracted via different   Arsenobetaine is reportedly the major arsenical in the Antarctic krill, with a concentration of 1.9 μg/g, corresponding to approximately 45% of the extracted arsenic (Grotti, Soggia, Goessler, Findenig, & Francesconi, 2010). As shown in Figure 5a,b, in the EH F I G U R E 5 Arsenic (a) and fluorine (b)  As shown in Figure 6, through the OPLS-DA analysis of the components in the KO, the KO of ethanol layer and that of hexane layer showed the largest difference, which proves the great separation effect of the EH method. In addition, the quality of KO extracted by subcritical butane-dimethyl ether was similar to that of the ethanol layer, and the quality of KO extracted by subcritical butane was close to that of the hexane layer. The KO extracted through Folch method was more comprehensive.
In general, the quality of KO in the ethanol layer was competitive, but its high water content and FFA content would affect its shelf life. Further study should be conducted on how to reduce the water content of the KO in the ethanol layer.

| CON CLUS IONS
In this study, KO was extracted via the EH method and subcritical

ACK N OWLED G M ENT
Financial support provided by National Natural Science Foundation of China; Marine Drugs and Biological Products (U606403) is appreciated.

CO N FLI C T O F I NTE R E S T
The authors declare that they do not have any conflict of interest.

E TH I C A L S TATEM ENT
This study does not involve any human or animal testing. Written informed consent was obtained from all study participants.