Supercritical fluid extraction (SFE) efficiency through response-surface optimization (RSO)

SFE extraction process was optimized based on the oil extraction performance. Results are shown in Table 1.

Table 1 Experimental results of the oil extraction performance of Spirulina sp. after supercritical fluid extraction (SFE)

Briefly, an increased oil yield was observed with the elapse of extraction time: in the 30-min treatments, oil yields of 2.27%, 1.75%, 3.28%, and 2.24% were obtained (Table 1, entries 6 and 10–12, respectively), while 75-min treatments showed oil yields ranging from 3.99% to 5.26%. Finally, after 120 min, the values of oil extracted were found between 4.27% and 6.25%.

In this point, it should be remarked that albeit a kinetic study would have been the most ideal solution to determine the optimal extraction time, this step was not considered in this study due to various reasons. Firstly, because of the nature of the material: microalgae possess a strong cell wall that sometimes requires harsher conditions to be broken down, for instance, high temperatures, pressure, or processing time. For that, we decided to operate in an acceptable range of conditions for these variables defined in the RSO in view of enhancing process efficiency, considering as the starting point our previous study [15]. Thus, pressure was applied from 20 to 40 MPa, the maximum value allowed by our supercritical equipment; 40–60 °C were considered for temperature, principally trying to avoid the degradation of certain bioactive molecules. Finally, time was optimized from 30 to 120 min since when a process intensification technology (SFE) is used, excessive time above 120 min can result contradictory.

Concerning again the kinetic study, kinetic curve can significantly vary depending on the temperature and pressure employed, so that different conclusions could have been obtained.

Furthermore, the reproducibility of the model can be assured attending to the oil performance obtained in the central points of the design (experiments 3, 9, and 13, Table 1), with proximate yields of 4.98, 5.05, and 4.00%, respectively.

Once the RSO was executed (15 experiments), the data recovered (i.e., oil performance) was again introduced in the Statgraphics Centurion XV® software, that provides the optimal extraction conditions derived from this model, as well as a thorough ANOVA analysis with regression coefficients for the optimized variable (see Supporting Material, section 2).

According to the extraction yields obtained, the optimal conditions for this design were 21.37 MPa, 120 min, and 60 °C, as displayed in Table 2 and Fig. 1. Figure 1a also shows these optimal conditions in the form of a graphical representation. It shows how the extraction yields increased with the elapse of extraction time, and that the change in pressure hardly influenced these results, slightly noting the maximum yield at 20 MPa.

Table 2 Optimal response for Spirulina with supercritical fluid extraction (SFE) treatment through response-surface modelFig. 1

Optimal conditions estimated by the response-surface design (a) and main effects of pressure, treatment time, and temperature on oil extraction performance (b)

The same is also shown in Fig. 1b, where the influence of the different variables (pressure, time and temperature) to recover the oil extraction performance is evident.

More specifically, it is observed that the pressure has the maximum performance at 21.37 MPa, and an increase in pressure slightly decreases the extraction performance. On the other hand, an increase in the duration of SFE treatment leads to a significant increase in performance, with 120 min being the optimal duration for this study. Finally, the extraction yield increases as the temperature increases, so that the highest results of this variable are obtained at 60 °C.

It is estimated that at these conditions, a predicted oil performance of 6.5% should be obtained (according to the data provided by Statgraphics Centurion XV®). Nonetheless, the execution of a SFE in these conditions to validate the RS model led to an oil performance of 5.03%, which corresponds with a defatting yield of 86.75% (Table 2). It can be thus concluded that there is no significant difference between the average of these yields and the expected optimal value. These results agree with those reported previously in the available literature, which states that the pressure of 4000 psi (~ 27.5 MPa) is the one that allows a higher extraction yield for two different species of microalgae: Arthrospira sp. and Schizochytrium sp. In the same study, cosolvent was also used to enhance scCO2 extraction, although methanol was used instead of ethanol [20]. These authors also observed an increase in lipid extraction with the elapse of the extraction time, and after 6 h of extraction (the maximum time spent), the maximum yield was obtained [20].

In respect to oil extraction yields obtained by the conventional Soxhlet treatment for 2 h, significantly lower outcomes were reported compared to those obtained via SFE during that time (0.53% oil yield, 9.18% defatting). This difference is clearly observed in the stains of the extracts obtained with SFE and the Soxhlet extractions (Fig. 2).

Fig. 2

Difference in the colour of the extracted oil. A Supercritical fluid extraction (20 MPa/60 °C/120 min). B Soxhlet 120 min

Taking into account the reviewed literature, Soalana et al. [21] obtained a superior performance in Soxhlet extraction (29%) versus recovery of oil extracted by SFE (~ 25%). However, it is worth mentioning that in that study, a difference of 16.5 h between the extraction time of the different methods was stated, with the Soxhlet being the one that was maintained for the longest time (18 h), while the SFE extraction was performed in 1.5 h. This difference in treatment conditions shows that Soxhlet requires too long extraction times, which has a greater economic and environmental impact.

Oil analysis

SFE allows the recovery of two extraction products. Most of the literature is focused on the analysis of the oil extract, with different assays for antioxidant activity, lipidic profile, etc. Herein, the main target is the post-extraction solid and its potential as a source of proteins and minerals, as well as bioactive peptides. Nonetheless, oil quality was also assessed and contrasted with some works reported in the literature.

Determination of antioxidant compounds

The total polyphenol content (TPC) and antioxidant activity (TEAC) obtained in the oil extracted from Arthrospira platensis by SFE and Soxhlet (conventional) are shown in Fig. 3. Regarding the total antioxidant capacity of the oil extracted by the TEAC assay, as shown in Fig. 3a, a large difference in TEAC values can also be observed between SFE (10.681 μmol TE/g DW) and the conventional Soxhlet extraction (0.987 μmol TE/g DW), with a significant difference between treatments once again (p < 0.0001). With the SFE treatment, it was possible to increase the extraction of antioxidant compounds 10 times compared to conventional treatment. Dejsunkranont et al. [22] reported a higher antioxidant capacity by performing several supercritical fluid extraction (SFE) experiments on a Spirulina maxima sample. These authors observed an antioxidant activity of 118.62 μmol TE/g under conditions of 31 MPa, 60 °C, and 120 min. However, in this case, the extraction was carried out without a cosolvent [22].

Fig. 3

Comparison of TEAC (Trolox Equivalent Antioxidant Capacity) (a) and TPC (total polyphenol content) (b) analyses in oil extracted from Spirulina sp. by supercritical fluid extraction (SFE) and Soxhlet as a control. Results expressed as mean ± SD. p < 0.0001. μmol TE/g DW, μmol Trolox equivalents per gram dry weight; mg GAE/g DW, mg gallic acid equivalents per gram dry weight

As can be observed in Fig. 3b, the value of TPC is clearly influenced by SFE treatment (4.600 mg GAE/g DW), which is significantly higher (p < 0.0001) with respect to the control (i.e., the conventional Soxhlet extraction) (0.156 mg GAE/g DW.) That is logical since it is not expected a high concentration of phenolic compounds in a crude recovered through a low polar environment (i.e., pure hexane). Therefore, it is shown how that innovative technologies, such as SFE, are considered more promising than conventional methods in terms of the recovery of bioactive compounds in microalgae [23].

It has been shown that the use of polar solvents favors the extraction of certain phenolic compounds. In the present study, ethanol was used as a cosolvent trying to maximize process efficiency and recovery of bioactive compounds, using the optimal conditions related to the highest extraction yield. A 10% stream is considered for the cosolvent based on our previous work [15] with no optimization addressed in this study.

Outcomes derived from this work can be contrasted with the study by Georgiopoulou et al. [23], who obtained the highest phenolic content in Chorella vulgaris by SFE + 10% ethanol (17.3 mg GAE/g extr.) compared to a conventional solid–liquid extraction (SLE) (11.02 mg GAE/g extr.). This difference is probably due to the recovery of phenolic compounds of different polarities. The results obtained in the present study are also related to the study by Wang et al. [24], in which microalgae, including Spirulina, were subjected to pressurized liquid extraction (PLE) and whose TPC is 12.5 mg GAE/g DW. In any case, the protocol herein applied led to a significantly lower TPC compared to these two works.

Determination of chlorophyll A and carotenoids

Spirulina contains distinctive natural pigments of orange, green, and blue (carotenoids, chlorophylls, and phycocyanins, respectively). In the present work, chlorophyll A (Chla) and carotenoids (Car) were analyzed (Fig. 4). Chlorophyll B has not been calculated in this study since, according to Lima et al. [19], the microalgae Spirulina lacks Chlb, and chlorophyll C is only present in brown algae.

Fig. 4

Content of chlorophylls and total carotenoids obtained from the microalgae Spirulina for supercritical fluid extraction (SFE) and conventional (Soxhlet) extraction. Results expressed as Mean ± SD. p < 0.0001. μg/g DW, μg per gram dry weight

Chlorophyll A, apart from the role it plays in photosynthesis, has demonstrated a strong antioxidant capacity [4]. In addition, it can be used as a natural colouring for some foods. On the other hand, carotenoids also contribute to the elimination of free radicals and are precursors of vitamin A [25].

For SFE, an increase in chlorophyll A extraction was observed compared to soxhlet. On average, a total of 114.985 μg Chla/g DW was obtained in the conventional Soxhlet extraction compared to 2650.484 μg Chla/g DW when applying the SFE extraction (i.e., up to 20 times higher). Statistical analysis indicates that the results obtained for chlorophyll A were statistically significant (p < 0.0001). This difference is also observable in Fig. 2, where the SFE extraction has a greener coloration than the conventional one, due to the greater presence of chlorophyll A.

These values can be contrasted with the work of Tong et al. [26], who focused on the extraction of chlorophyll A from the microalgae Arthrospira platensis by SFE; therefore, the conditions were optimized according to chlorophyll A. The optimal conditions used were 40 MPa, 323.2 K (around 50 °C), 1 h of treatment duration, 8 g/min of CO2 flux, and 10 mL of different cosolvents, including ethanol. The results of chlorophyll A obtained were almost 2.5 mg/g. Moreover, compared to conventional treatment, about 3 times more chlorophyll A was extracted during the same extraction time by these authors. In addition, it was observed that the presence of a cosolvent facilitates the extraction of chlorophylls, since the treatment with scCO2 without any cosolvent was also analyzed and hardly any chlorophylls were extracted [26].

On the other hand, SFE treatment also significantly increased carotenoid extraction. On average, 532.430 μg Car/g DW was extracted after SFE and 76.061 μg Car/g DW after conventional extraction, so that SFE treatment increased carotenoid extraction by 7 times compared to soxhlet extraction, which was statistically significant (p < 0.0001).

These results can be compared with the study by Marzorati et al. [4], in which scCO2 was used for carotenoid extraction under the following conditions: 300 bar (30 MPa), CO2 flux of 15 mL/min, 45 °C, and 1 h of extraction (without any organic cosolvent, which is more selective for carotenoid extraction). These authors obtained 3.5 mg of Car/g DW in the extract obtained under the above conditions.

Determination of minerals

Essential minerals are vital for maintaining various bodily functions, including building materials for human bones, muscle contraction, nerve function, and regulating the body’s water balance. They are also components of hormones, enzymes, and other biologically active compounds. Some minerals also play an important role in the optimal functioning of the immune system [27]. Spirulina contains a considerable amount of potassium (K). Moreover, it provides major minerals (calcium (Ca), phosphorus (P), and magnesium (Mg)) at ratios similar to those present in milk. It is also important to note that Spirulina is an exceptional source of iron (Fe), containing approximately ten times more than other iron-rich foods [28, 29].

Concerning mineral content in the Spirulina extracts obtained via SFE and Soxhlet (Fig. 5), significantly higher values were found in the SFE extracts for Mg (5.3 vs 0.21 mg/L), P (1.3 vs 0.5 mg/L), K (4.1 vs 0.8 mg/L), and Fe (122 vs 39.6 µg/L), being the Mg and K contents of the SFE extracts remarkable. However, non-significant differences were observed for Ca (0.9 vs 1.25 mg/L), Cu (8.2 vs 16.4 µg/L), and Zn (83.3 vs 86.6 µg/L) between the SFE and Soxhlet extracts.

Fig. 5

Major (Mg, P, K, Ca) and minor (Fe, Cu, Zn) minerals present in the recovered oils after SFE and conventional extraction of spirulina

In any case, it is important to note that in all cases the mineral content of the SFE extracts was higher compared to conventional extraction, highlighting the role of SFE technology in the extraction of essential minerals. The same trend was observed in a previous study where SFE improved mineral extraction, increasing from 49.33 to 91.50 µg/g DM for Mg and from 0.87 to 2.00 µg/g DM for Fe compared to conventional extraction, while conventional was more effective in recovering P and Ca [15]. Regarding other innovative extraction technologies, pressurized liquid extraction (PLE) allows an extraction rate close to 100% for Zn, while Mg, Ca, and Fe were extracted from Phaeodactylum tricornutum at rates of 12.4%, 0.6%, and 4.8%, respectively [15, 24].

Assessment of nutritional properties of solid residues: protein content, bioactive peptides, and mineral profile

As has been previously referred, the analysis of spirulina residues (i.e., post-extraction solid that remains with no further utilisation) for further upcycling is the main attainment in this work with respect to most of the literature reporting SFE of spirulina, in which this solid is treated as a waste, with no considerations for its recycling and any previous analysis. In this regard, preliminary assessment of this material addressed to determine its nutritional profile, namely protein content, bioactive peptides, and mineral profile.

Protein content and bioactive peptides

Outcomes derived from the total protein content analysis in raw Spirulina, as well as in the cakes obtained by Soxhlet and SFE methods, are shown in Table 3.

Table 3 Effect of supercritical fluid extraction (SFE) on Spirulina proteins

No significant differences were observed in the protein contents of untreated microalgae samples (69.2 g/100 g) and the remaining cakes obtained after conventional Soxhlet extraction, while SFE-treated Spirulina samples showed a slightly increased (74.4 g/100 g). That can be explained since a similar extraction time was applied to both procedures (120 min) to process the material. Therefore, the intensification mechanisms involved in the supercritical CO2 technology (i.e., high pressures, higher diffusivity, solubilization, etc.) would facilitate a more efficient extraction of oil than conventional systems, thus leading to a higher enhancement in the post-extraction recovered material, as has been observed.

Similar results were reported in the study conducted by Qiuhui et al. [30], in which the different nutrients present in Spirulina were evaluated by carrying out a control treatment and another with scCO2. The authors obtained 69.75 g protein/100 g in the control, while the protein values found after scCO2 treatment were 69.79 g/100 g.

Moreover, a relevant analysis that involves a tentative identification of antioxidant peptides was also performed, once again focused on the remaining cakes obtained by SFE and Soxhlet extraction (Tables 4 and 5).

Table 4 Antioxidant peptides found in the Spirulina remaining cake after supercritical fluid extraction (SFE) with potential antioxidant capacity and the group responsible for this bioactivityTable 5 Antioxidant peptides found in the Spirulina remaining cake after Soxhlet extraction with potential antioxidant capacity and the group responsible for this bioactivity

For this purpose, the bioactivity and potential bioactivity of the identified peptides were checked using the BIOPEP-UWM database. Only peptides with a score of more than 90% were considered. The database did not show any correlation between the identified peptides and antioxidant activity. However, the presence of several short amino acid sequences with antioxidant properties within these peptides was revealed. As can be observed in Tables 4 and 5, 14 peptide sequences were found including specific amino acids with potential antioxidant capacity in SFE residue cakes compared to the 9 observed in Soxhlet remaining cakes. The following peptide sequences including specific groups with potential antioxidant activity “AYNGDPFVGHLSTPISDSAFTR” (HL, AY), “MKTPLTEAVSIADSQGRFLSST” (LT), “NDAAGDGTTATVLAHAMVKE” (AH), “NGDPFVGHLSTPISDSAFTR” (HL), and “TSKADSLISGAAQAVYNKFP” (VY, GAA, GA) were tentatively identified in both SFE and Soxhlet remaining cakes. It is important to note that the peptide sequences “SIVNADAEARYLSPGELDRIK” and “RYLSPGELDRIKSFVT” were only detected in the remaining SFE cakes, having a significant antioxidant potential, with 4 short amino acid groups associated with potential antioxidant capacity, in both cases “EL”, “RY”, “RYL”, and “YL”.

Previous studies performed in algae biomass have reported that peptides containing hydrophobic amino acids (Leu, Val, Trp, Ala, Phe, and Tyr) in their structure could exhibit higher free radical scavenging activity by helping the peptides to pass through membrane lipid bilayers. In addition, acidic or basic amino acids (Glu, Asp, and Lys) could act as metal ion chelators, while aromatic amino acids (Tyr, Trp, and Phe) could quench free radicals by electron transfer [31, 32]. Nonetheless, as far as our knowledge goes, this is the first study that tackles a tentative analysis of the bioactive peptides present in the spirulina SFE residues, which points out the value of this work.

With this tentative identification, the potential of spirulina residues as antioxidant materials is corroborated, albeit further studies or investigations would be required in order to assess the real antioxidant power of these materials.

Mineral profile

In respect to minerals’ content evaluated in untreated spirulina and remaining cakes obtained after SFE extraction and Soxhlet (Fig. 6), the major minerals (Ca, P, Mg) were detected without significant differences in the raw samples and the remaining cakes, with values ranging from 3.9 to 4.2 mg/g for Mg, 10.5 to 10.7 mg/g for P, and 2.7 to 2.8 mg/g for Ca. Potassium was the mineral quantified at higher levels, with a significant difference between SFE (12.83 mg/g) and Soxhlet (12.03 mg/g) residual cakes. Regarding minor minerals, no significant differences were observed for Cu and Zn, which were found in the raw material and remaining cakes at levels of 0.28–0.51 mg/kg and 9.3–11 mg/kg, respectively. The mineral Fe was detected in an interesting amount between 838 and 861 mg/kg. In this case, the level was significantly lower in the SFE residual cake compared to the Soxhlet cake; however, as has been mentioned before, a higher amount of Fe was detected in SFE extract compared to the conventional Soxhlet extract (122 µg/L vs 39.6 µg/L). A previous study reported similar levels of Mg and P in Spirulina (3.2 mg/g dw) and (9.3 mg/g), respectively. However, these authors reported higher levels of Zn (44 mg/kg) [24]. The results obtained in the present study suggest that most of the minerals remain in the cakes after extraction with SFE and conventional Soxhlet extraction; however, the extracts obtained with SFE are richer in all the minerals analysed in this study compared to conventional extracts, highlighting their Mg, K, and Fe contents.

Fig. 6

Major (Mg, P, K, Ca) and minor (Fe, Cu, Zn) minerals present in the untreated material and the remaining cakes obtained after SFE and conventional extraction