3.6. SFE–US with cosolventsTable 3(C) shows that the addition of ethanol and water in SFE-US from the dried and milled material leads to higher extractionyields. This behavior is related to the polar nature of the chosencosolvents, which allows extracting components that would notbe soluble in pure CO2, which is non polar. The molecules of thecosolvent compete with the active sites of the substrate to inter-act with the extractable compounds. Therefore, the presence ofcosolvents helps breaking the interactions between substrate andsolute, which can be solubilized by the solvent mixture [7,38]. Theincrease of solubility caused by a cosolvent results from the for-mation of cosolvent–solute and solvent–cosolvent–solute groups[39]. A cosolvent with critical temperature lower than that of thesupercritical fluid usually reduces the solubility of low volatilitycompounds, and the opposite occurs if the critical temperatureof the cosolvent is higher [5]. Both cosolvents used in this work(ethanol and water) have critical temperatures higher than CO2(240.6◦C, 374.2◦C and 31.1◦C, respectively), so they enhance thesolubility of many compounds present in the blackberry bagasse.An increase in SFE-US yield is observed when ethanol was usedas cosolvent, compared to water. Besides increasing yield, ethanolcan be removed easier than water, so its application is widelyreported [40]. This behavior can be explained by the enhanced solu-bility of polar compounds in the mixture CO2+ ethanol. Moreover,the use of ethanol as cosolvent may have increased the numberof extracted compounds, thus reducing the selectivity of the pro-cess. Although water is more polar than ethanol, the SFE yields withwater as cosolvent were lower. Polarity is not the unique factoraffecting extraction yield. The type of interactions between solventand solute should also be comprehended.
As observed for SFE from blackberry bagasse, other works reportremarkable enhancements in SFE yield by using ethanol as cosol-vent at low concentrations. Luengthanaphol [41] compared yieldand antioxidant activity of extracts (Tamarindus indica L.) fromtamarind seeds obtained by SFE with pure CO2and with 10%ethanol as cosolvent, and verified that SFE of antioxidants is sig-nificantly improved with ethanol. Kitzberger et al. [42] observedan increase in the SFE yield from shiitake from 1.01% to 3.81% byusing 15% ethanol as cosolventThe contribution of ethanol and water as cosolvents in the recov-ery of phenolics, anthocyanins, and the antioxidant activity of theextracts can be observed in Fig. 3. In general, both cosolvents hadpositive influence on the extraction of the mentioned compounds,and the effect of water as cosolvent was clearly higher than ethanol.In SFE-US with cosolvents the highest anthocyanin content(17.54 ± 0.07 mg cyanidin 3-O-glucoside/g extract) was obtainedusing 5% water for the fresh sample. Tena et al. [43] and Murga et al.[44] reported that adding cosolvent to CO2helps improving theextraction yield of some compounds, such as anthocyanins, phen-olics, and antioxidant activity. It can be noted in Table 3(C) thatthe cosolvent ratio affects the anthocyanin concentration in theextracts, indicating that interactions between solute and substratemust have been broken and replaced by cosolvent molecules [7].Thus, the anthocyanin yield increased with the cosolvent concen-tration [45]. This is clear in the extractions from dried and crushedsamples, where raising the water ratio from 5 to 10% stronglyincreased the anthocyanin concentration, due to the enhancementof the solute/cosolvent interactions that raise solubility [7]. Thesame effect is also noted in the extractions with ethanol, althoughthe anthocyanin recovery was lower than with water. Thus, wateras cosolvent not only increases SFE yield, but is also the most ade-quate solvent since is ecologically safe and cheap.Regarding phenolic compounds, the best results of SFE-US werefound using 10% water in the dried and crushed sample and 5%water in the fresh sample. As expected, water extracts phenolicsefficiently, and the extracts with most phenolics also presentedthe highest antioxidant activities measured by DPPH and ABTS,evidencing some correlation between phenolics and antiradicalcapacity [46,47]. However, the relation shows that phenolics arenot the unique responsible for the antioxidant activity of theextracts. Vegetable substrates contain several phenolic compo-nents with different antioxidant activities, and the synergismbetween antioxidants in a mixture makes their activity dependentof concentrations, structure and their chemical interactions [48].It can also be observed in Table 3(C) that the method ABTS ismore effective in the detection of antioxidant compounds thanDPPH. The method DPPH is widely applied to determine the antioxi-dant activity in extracts and isolated compounds, such as phenolics,anthocyanins, flavonols and cumarins [49,50]. Every method pro-vides precise and reproducible results, but the antioxidant activitiesmay differ significantly from one method to other. Thus, mostmethods provide partial results regarding antioxidant activity ofcomplex extracts [51,52].The addition of water as cosolvent at 10% provides the highestantioxidant activities by both methods. For the extracts obtained bySFE-US with pure CO2the antioxidant activities were quite lower,indicating again that this property is intimately related to pheno-lics, which are polar compounds that can hardly be extracted witha nonpolar solvent like CO2.Finally, evaluating the extracts obtained by SFE-US with cosol-vent, one can conclude that to achieve the best yields it is preferableto use 10% ethanol, which provided yields eight times higher thanwith pure CO2. Since ethanol is a slightly polar solvent, its addi-tion as cosolvent allowed the dissolution of polar substances thatwere not extracted with pure CO2. In terms of anthocyanins, phen-olics, and antioxidant activity, it is better to use 10% water ascosolvent. The increase of such compounds may be due to thelow solubility of water in CO2, which may lead to the coexistenceof two phases. In this case, a liquid phase containing water asmajor component would help extracting preferentially phenolicsand anthocyanins. Nevertheless, if SFE is performed on the freshsample, 5% water is more recommendable, since the water con-tent of the sample seems to work as cosolvent and enhance theextraction of anthocyanins. Moreover, the recovery of anthocyaninscould have been enhanced due to pH reduction in the presence ofCO2and water, since anthocyanins are usually more stable in acidi-fied media [53–55]. Summarizing, the differences between SFE-USwith and without cosolvents can be attributed to the changes in cosolvent concentration, type of pretreatment, and extractionmethod [56].