3.1.2. Flavonoids

With regard to flavonoids and its several sub-classes, only flavones and their derivatives were detected in artichokes by-products. Thereby, peak 6, with *m/z* at 593.1512, was tentatively identified as luteolin-rutinoside [18], while peak 7 with *m/z* at 447.0933 was proposed as luteolin-glucoside (cymaroside) according to bibliographic data [10,21]. Furthermore, the compound eluting at 8.58 min (peak 9) and displaying *m/z* 577.1563 was considered as apigenin-rutinoside (isorhoifolin) [21]. In the same way, peak 10, detected at 8.98 min and *m/z* 431.0984, was tentatively assigned to apigenin-glucoside according to the comparison of the molecular formula provided by the detector and literature data [21]. Similarly, peak 13 at *m/z* 285.0405 and molecular formula C15H10O6 along with peak 14 at 15.44 min and *m/z* 269.0455 were proposed according to the literature as luteolin [10,21] and

apigenin [10,21], respectively. Finally, peak 17 with *m/z* 283.0612 was tentatively assigned to methylapigenin [22].

**Figure 2.** Extracted ion chromatograms (EIC) and mass spectra of some phenolic compounds characterized by HPLC-time-of-flight mass spectrometry with an electrospray interface (ESI-TOF/MS) in artichoke by-products.

### 3.1.3. Saponins, Lipids, and Other Polar Compounds

Peak 1, with *m/z* at 191.0561, was proposed as quinic acid, a carboxylic acid commonly found in artichoke [21].

On the other hand, compounds belonging to saponins family correspond to peaks 18 and 19. They displayed a *m/z* 925.4802 and retention times of 26.62 and 26.92 min, respectively. Both were identified as isomers of cynarasaponin A or H [21].

Lastly, several lipids were also found in these artichoke samples. Thence, hydroxyoctadecatrienoic acid (peak 21) is a lipid previously identified in artichoke [21]. Trihydroxyoctadecenoic acid (peak 15), hydroxy-octadecadienoic acid (peak 22), and linolenic acid (peak 23) were also detected in the present analysis. All these lipids have been previously identified in other vegetable PLE extracts, such as *Morus nigra* [23] and *Prunus avium* [24]. Similarly, dihydroxyhexadecanoic acid (peak 16) was identified in the present study as well as in *Symphytum officinale* L. samples [25]. Although these plant matrices belong to different species, both molecular formula and *m/z* generated by the software match with data reported in the bibliography.

### *3.2. Extraction Yield and Extraction Efficiency*

In this study, different PLE conditions were evaluated and compared in terms of yield and recovery of individual phenolic compounds from artichoke by-products. The extraction yield was expressed as weight of collected extract per dry plant material (*w*/*<sup>w</sup>*, in grams) used in the extraction procedure. The extraction efficiency of individual compounds was estimated by a semi-quantitative approach, measuring the peak areas of the compounds identified in the chromatogram as expressed as mean ± standard deviation of three consecutive injections. Table 3 includes the extraction yield and phenolic extraction efficiency for the *C. scolymus* L. PLE extracts.


**Table 3.** Extraction yield and phenolic compounds recovery for *C. scolymus* L. obtained in each PLE condition. Yield (%), individual phenolic compounds (peak area × E + 4). Value = mean ± standard deviation.

> Concerning the extraction yield, the obtained results in the different PLE experiments are highly inconstant. Similar variations were observed in other PLE extracts from plants such as black mulberry or sweet cherry stems, ranging from 11% to 48% or from 3% to 49%, respectively [23,25]. In general, it could be observed that the application of elevated temperatures (above 170 ◦C) resulted in higher extraction yields. Indeed, PLE 3 (200 ◦C, EtOH 50%), PLE 2 (176 ◦C, EtOH 85%), and PLE 5 (176 ◦C, EtOH 15%) were the extraction conditions with the highest yield values (57 ± 2%, 50 ± 1%, and 45 ± 2%, respectively). This fact could be explained by the increase in solvent diffusivity with increasing temperature, which enhances the extraction of several components from vegetable matrices [26]. In contrast to this observation, it could not establish a relationship between the percentage of ethanol and extraction yield.

Focusing on the individual recovery of phenolic compounds of each PLE condition, a relationship between the extraction parameters (temperature and percentage of ethanol) and the enhancement extraction of particular compounds could be revealed.

In this sense, regarding phenolic compounds, as described above, both sub-classes of polyphenols (phenolic acids and flavonoids) identified in the samples were extracted under all the PLE experiments. The individual areas of all compounds identified in PLE extracts were also analyzed to evaluate significant differences among the different PLE conditions. Figure 3 draws the peak area (mean value ± standard deviation) of the individual compounds detected in each artichoke PLE extract as well as the total phenolic acids and flavonoids (see Table S1).

**Figure 3.** Abundance of phenolic compounds characterized in each *C. scolymus* PLE extracts by HPLC-ESI-TOF-MS: (**A**) individual phenolic acids, (**B**) total phenolic acids, (**C**) individual flavonoids, and (**D**) total flavonoids.

In regards to the total abundance of phenolic acids in the *C. scolymus* PLE extracts, it could be observed that PLE 1 (120 ◦C, EtOH 100%), PLE 4 (63 ◦C, EtOH 85%), and PLE 7 (40 ◦C, EtOH 50%) conditions showed the best recovery for this chemical sub-class. On the contrary, PLE 5 (176 ◦C, EtOH 15%) and PLE 2 (176 ◦C, EtOH 85%) present the worst extraction recoveries for this kind of substances. In light of these data, it can be concluded that higher percentages of ethanol in aqueous mixtures and moderate to high extraction temperature enhances the extraction of these compounds. On the contrary, very high temperatures, such as in PLE 5 and PLE 2, seem to act to the detriment of the extraction of phenolic acids.

Additionally, the individual behavior of phenolic acids was also monitored. Chlorogenic acid was extracted in greater abundance by PLE 1 condition (with the highest percentage of ethanol, 100%), followed by PLE 4 (63 ◦C, EtOH 85%) and PLE 7 (40 ◦C, EtOH 50%) experiments, which presented similar abundances. In the case of rosamarinic acid, the best extraction conditions were PLE 7 and PLE 4. Both experiments were performed with low-intermediate temperatures and medium-high percentages of ethanol. In fact, higher temperatures such as the applied in PLE 5 (176 ◦C, EtOH 15%) recovered this compound in very low amount or could not extract rosamarinic acid from artichoke by-products, as in PLE 2 (176 ◦C, EtOH 85%) and PLE 3 (200 ◦C, EtOH 50%). Finally, within the family of phenolic acids, cynarin isomers showed different behavior over the tested extraction conditions. PLE 1 run (120 ◦C, EtOH 100%) reported a high value of cynarin isomer 1 peak area, while cynarin isomer 2 was not recovered in abundance. The same phenomena could be described for PLE 4 (63 ◦C, EtOH 85%) and PLE 7 (40 ◦C, EtOH 50%).

On the other hand, the analysis of the total abundance of flavonoids in the artichoke extracts indicated that PLE 1 (120 ◦C, EtOH 100%) was the best condition, followed by PLE 4 (63 ◦C, EtOH 85%) and PLE 7 (40 ◦C, EtOH 50%). All these runs applied intermediate conditions within the range of temperature and percentage of ethanol as solvent combinations. On the contrary, the conditions PLE 2 (176 ◦C, EtOH 85%), PLE 5 (176 ◦C, EtOH 15%), PLE 8 (120 ◦C, EtOH 0%), and PLE 9 (63 ◦C, EtOH 15%) reported a lower abundance with respect to the other PLE extracts.

Analyzing individual flavonoids, the highest abundance of glycoside structures was obtained under the conditions PLE 1 (120 ◦C, EtOH 100%) and PLE 4 (63 ◦C, EtOH 85%). On further consideration, aglycon flavonoids as luteolin and apigenin showed the same trend, being better extracted by the PLE 1 condition. The abundance of simple flavonoids, luteolin and apigenin, was not increased as extraction temperatures raised. This fact suggests that high temperatures did not generate the hydrolysis of glycoside forms over the extraction process (luteolin-glucoside, apigenin-glucoside, luteolin-rutinoside, and apigenin-rutinoside).

Taking into account all these results, a comparison of the extraction yield and phenolic compounds recovery in each experimental PLE run showed important differences. Although the high temperatures would improve the diffusivity of the solvent and break component–matrix interactions, which increase the solubility of the analytes and consequently enhance the extraction yields [27], it does not necessarily mean a greater recovery of phenolic compounds. This fact seems to occur with artichoke by-products, which could be related to different factors. Indeed, an excessive increase in temperature is demonstrated to negatively affect the extraction of thermolabile compounds, such as phenolic compounds [26].

In addition to the possible thermal degradation of compounds, different combinations of solvent composition and temperature provide changes in the dielectric constant value, which is crucial for the extraction recovery of the phytochemicals from natural sources. In this way, special attention has to be paid to PLE 1 condition (120 ◦C, 100% EtOH) with a dielectric constant value of 19. Despite this condition reported the lowest yield, the recovery of phenolic compounds was higher compared to other PLE conditions. This phenomenon has also been reported in other plant matrices extracted by PLE at the same conditions (100% EtOH and 120 ◦C) [23].

Thereby, PLE 1 condition provided 2.6, 7.6, and 3.4 times higher recovery of phenolic acids, flavonoids, and total phenolic compounds than PLE 8 (120 ◦C, EtOH 0%, dielectric constant 50.4), respectively. Analyzing individual phenolic compounds identified in both PLE conditions, significant differences could be observed concerning the abundance of all of them (see Table S2). A similar trend was observed when the yield and recovery of phytochemical compounds of PLE 6 (120 ◦C, EtOH 50%, dielectric constant 34.7) were compared to those obtained for PLE 1 (120 ◦C, EtOH 100%, dielectric constant 19.0). Thus, the yield obtained for PLE 6 is 5.4 times higher than for PLE 1. Nevertheless, PLE 1 condition showed 2.6 times higher phenolic recovery than PLE 6.

Nonetheless, tt is important to remark that at dielectric constant values higher than 19, minor differences were observed on the yield and recovery of phenolic compounds among the experimental conditions. Indeed, PLE 8 (dielectric constant value, 50.4) obtained an extraction yield 1.1 times higher than PLE 6 (dielectric constant value, 34.7). Moreover, PLE 6 had 1.1, 2.2, and 1.3 times higher recovery of phenolic acids, flavonoids, and total phenolic compounds than PLE 8, respectively, significant differences among all the characterized phenolic compounds were found for PLE 8 and PLE 6, except for chlorogenic and rosamarinic acids (Table S2).

Therefore, similar differences on the yield and phytochemical recovery were also observed for PLE conditions where the same temperature was applied in combination with different percentages of ethanol, which consequently provides different values of dielectric constants. Some examples are: (a) temperature of 63 ◦C combined with 15% of ethanol (PLE 9, dielectric constant value 59.1) or 85% of ethanol (PLE 4, dielectric constant 31) and (b) extractions at 176 ◦C using 15% EtOH (PLE 5, dielectric constant 33.4) or 85% EtOH (PLE 2, dielectric constant 21.6). These results pointed out that at the same working temperature, the recovery of phenolic compounds is improved when the solvent composition (% EtOH) provides the lowest value of dielectric constant.

Nevertheless, despite this focus in the targeted extraction of individual compounds provided by concrete PLE parameters, the total extraction yield should be considered if the main purpose is to obtain the maximum quantity of extract with a particular compound.
