Screening of Volatile Compounds in Mate ( Ilex paraguariensis ) Tea—Brazilian Chimarr ã o Type—By HS-SPDE and Hydrodistillation Coupled to GC-MS

: The volatile fraction of mate ( Ilex paraguariensis ) tea—speciﬁcally Brazilian chimarr ã o type, which has an odor proﬁle comprising distinctive fresh, green, grass, and herbal notes—was investigated. Hydrodistillation in a Clevenger apparatus was employed in order to extract volatiles from the tea matrix. Headspace–solid-phase dynamic extraction (HS-SPDE) was employed to extract the volatiles from two types of infusions of this tea—a simple single infusion and a traditional preparation of consecutive infusions. Volatiles were analyzed by gas chromatography–ﬂame ionization detection/mass spectrometry (GC-FID/MS). In total, 85 compounds were either identiﬁed or tentatively identiﬁed and semi-quantiﬁed. Semi-quantiﬁcation comprised peak area integration of all the peaks (including the unidentiﬁed ones) in the chromatogram. Results obtained by hydrodistillation and by HS-SPDE were distinct, covering mostly different ranges of volatility and showing only 15 compounds in common. The identiﬁed compounds had their respective average and minimum odor thresholds and odor characteristics compiled from the literature. Several major compounds considered as key odorants in other mate tea products were not detected or only present at low levels in the samples of this research. Approximately half of the odorants identiﬁed in these samples were commonly reported in different mate tea types; the remaining 41 molecules—predominantly terpenoids (isoprenoids)—could be listed as speciﬁc to the Brazilian chimarr ã o type and are suggested to underlie its typical freshness.


Analytical Approaches and Instrumentation
When assessing the contribution of volatiles to the flavor perception, some factors need to be taken into account. Different extraction techniques applied for the evaluation of mate teas have generated different results [3]. Some studies performed analyses of mate tea distillates [1,2,7,8]. Techniques based on headspace sampling and lower temperatures than those used for distillation were successfully applied for mate and other teas with good sensitivity and reproducibility [3,5,9,10]. Other studies about mate tea volatiles employed gas chromatography (GC)-for separation-and its common hyphenated techniques for detection, identification, quantification, and semi-quantification [1][2][3][4][5][6][7]. Semi-quantification has already been performed in other studies involving numerous mate tea volatiles by integrating all the peak areas registered in the chromatogram and calculating their individual relative peak area [1,2,5,8].
Finally, for the interpretation of analytical results it is necessary to consider that the potential sensory impacts of the different volatiles tend to be not just directly proportional to their concentrations but also inversely proportional to their odor thresholds-among several other factors [11].

Mate Tea Infusions: Traditional Consecutive Infusions and Single Infusions
An important feature of mate tea that must be taken into account is its most common way of consumption, in traditional consecutive infusions (TCI) in a gourd, popularly known as 'mate' (or also 'chimarrão' in Brazil). In this case, mate tea is poured into a gourd, accommodated on one side of it, and water is added to fill up the remaining empty space within the gourd. A metallic straw with a filter coupled to its lower end is then placed inside the gourd. The straw is slowly sucked and the infusion drunk until the gourd is empty. Once empty, more water is added to the gourd. This process is repeated several times. A gourd is commonly shared by several people in a communal way [12,13]. All these features differentiate these mate-tea-specific preparations from other herbal teas.
Even though mate tea is mostly consumed in the form of the traditional consecutive infusions described above, it is also used to prepare a simple single infusion (SI), as other common herbal teas, by infusing a small portion of tea under hot water during a certain extraction time [14,15]. Mate tea single water infusions were studied in other research, which employed or covered various tea-to-water ratios and water temperatures [13][14][15][16][17][18][19]. In some Spanish-speaking countries, this single infusion is called 'mate cocido'. It is prepared with boiling water and drunk very hot, hot, or warm. Nevertheless, consumption at an excessively high water temperature is correlated with a higher occurrence of esophageal cancer and thus is not advisable [18]. Nowadays, the preparation of a mate tea infusion 'in tea bag form is also common' [19].

Mate Tea Types
Mate beverages are widely consumed in southern Brazil, Argentina, Paraguay, and Uruguay [12]. As could be expected, consumer preferences vary notably among these countries. Therefore, as a result of specific methods of processing and aging employed in the different countries and companies, various mate teas of different types or qualities exist ( Figure 1) [14].

Aims of This Research
The aim of this work was to extend our knowledge of the potential key odorants present in Brazilian chimarrão mate teas and in their most common infusions by combining different approaches. These approaches were: identification and semi-quantification of the volatiles found in the essential oil and in the different water infusions (TCI and SI) of these teas; the definition of volatiles that are specific to this type of tea; and the appreciation of the odor thresholds of these molecules.

Aims of This Research
The aim of this work was to extend our knowledge of the potential key odorants present in Brazilian chimarrão mate teas and in their most common infusions by combining different approaches. These approaches were: identification and semi-quantification of the volatiles found in the essential oil and in the different water infusions (TCI and SI) of these teas; the definition of volatiles that are specific to this type of tea; and the appreciation of the odor thresholds of these molecules.

Samples
Samples of Brazilian mate tea-chimarrão-were produced following the traditional industrial process for this type of product. After blanching, drying, and grinding, mate teas were immediately packed in 1 kg vacuum packs and, without any aging period, transported to the laboratory in Germany and stored frozen (−20 °C) until analysis. The two samples were (as specified in the labels): one of Brazilian 'chimarrão tradicional premium' type produced in August (A) and another of Brazilian 'chimarrão tradicional' type produced in November (B). These samples were produced by Barão Comércio e Indústria de Erva Mate LTDA (Barão de Cotegipe, RS, Brazil) and analyzed as conventionally commercialized and consumed. No grinding processes were applied.

Hydrodistillation: Extraction of Volatiles in the Mate Tea Samples
Both mate tea samples (A and B) were subject to hydrodistillation. Hydrodistillation is based on the European Pharmacopeia 9.0 [20] and other studies [2,8]. A large sample of tea and a small volume of water had to be used in order to produce an appreciable sample of extracted essential oil. A total of 100 g of tea sample, 700 mL of distilled water, and ten boiling chips (IDL GmbH & Co KG, Nidderau, Germany) were added to a 2000 mL round bottom flask and the Clevenger apparatus was fitted on top of it. During the onset of boiling, an extra 100 mL of room temperature distilled water had to be slowly added from the central orifice of the Clevenger apparatus to control the initial foam formation in the neck of the round bottom flask. After two hours of distillation, 3 × 83.3 µL of n-hexane (Merck

Samples
Samples of Brazilian mate tea-chimarrão-were produced following the traditional industrial process for this type of product. After blanching, drying, and grinding, mate teas were immediately packed in 1 kg vacuum packs and, without any aging period, transported to the laboratory in Germany and stored frozen (−20 • C) until analysis. The two samples were (as specified in the labels): one of Brazilian 'chimarrão tradicional premium' type produced in August (A) and another of Brazilian 'chimarrão tradicional' type produced in November (B). These samples were produced by Barão Comércio e Indústria de Erva Mate LTDA (Barão de Cotegipe, RS, Brazil) and analyzed as conventionally commercialized and consumed. No grinding processes were applied.

Hydrodistillation: Extraction of Volatiles in the Mate Tea Samples
Both mate tea samples (A and B) were subject to hydrodistillation. Hydrodistillation is based on the European Pharmacopeia 9.0 [20] and other studies [2,8]. A large sample of tea and a small volume of water had to be used in order to produce an appreciable sample of extracted essential oil. A total of 100 g of tea sample, 700 mL of distilled water, and ten boiling chips (IDL GmbH & Co KG, Nidderau, Germany) were added to a 2000 mL round bottom flask and the Clevenger apparatus was fitted on top of it. During the onset of boiling, an extra 100 mL of room temperature distilled water had to be slowly added from the central orifice of the Clevenger apparatus to control the initial foam formation in the neck of the round bottom flask. After two hours of distillation, 3 × 83.3 µL of nhexane (Merck KGaA, Darmstadt, Germany) were used to flush the glass surfaces around the few droplets of essential oil extracted. This mixture was collected with 8 mL of the hydrosol (the aqueous phase obtained from the hydrodistillation) in a 10 mL test tube. The supernatant (non-polar phase) was collected with a pipette, transferred into a 1.5 mL vial, and immediately analyzed. Distillations were performed in triplicate for each tea sample and the mixture of essential oil and n-hexane was analyzed without further dilution prior to liquid injection.

Preparation of Popular Mate Tea Infusions: Single Infusion and Traditional Consecutive Infusions
Both mate tea samples (A and B) were also employed for a lab simulation of two different popular mate tea infusions: one to be representative of the preparation of a single infusion, such as for a conventional tea, and another to be representative of the preparation of traditional consecutive infusions in a gourd.
The conventional single infusion (SI) was accomplished by simply adding 3 g of tea and 200 mL of distilled water at 70 • C to a 250 mL beaker, simulating a domestic preparation of tea, similar to the approaches of other studies [13,16,17]. After one minute, the infusion was collected with a traditional 'bomba' or 'bombilla', a stainless-steel straw with a filter at its lower end (Bortonaggio, Garibaldi, RS, Brazil). The filter at the lower end had a diameter of 34 mm, with 160 holes (60 on each side) of 1 mm each. Over this stainless-steel filter, another finer filter (J.M. Filtros, José Luís Pereira & CIA. LTDA, São Leopoldo, RS, Brazil) with a pore size of around 200 µm was fitted. A 100 mL syringe was coupled with a silicon hose to the upper end of this metallic straw. After suction of the first 100 mL of infusion, 5 mL aliquots were added to the 20 mL SPDE vials. SI infusions were performed in triplicate for each tea type, followed by single headspace-solid-phase dynamic extraction (HS-SPDE) for each replicate.
Traditional consecutive infusions (TCI) ( Figure 2) were performed based on the procedures of Meinhardt et al. [13]. A homogeneous sample of 48 g of mate tea was added to a 223 mL glass gourd (Meta Mate, Berlin, Brazil) and agitated manually back and forth, in horizontal position. After shaking, the recipient was positioned at a 45 • angle and received the water for hydration (145 mL at 20 • C). After the hydration step of 5 min, the first cold infusion was sucked, and 9 consecutive infusions were performed. The consecutive infusions consisted of adding water up to the edge of the gourd, allowing 30 s of infusion time, and sucking the infusion with the syringe coupled to the upper tip of the metallic straw. Only the 10th infusion was then transferred to a 200 mL beaker and 5 mL aliquots were added to a 20 mL SPDE vial by pipetting. TCI infusions were performed in triplicate for each tea type, followed by single extraction by HS-SPDE for each replicate.

HS-SPDE: Extraction of Volatiles in Infusions
The parameters for HS-SPDE were based on previous studies [21,22]. Within a 20 mL SPDE glass vial, 5 mL of the infusions and 100 µL of internal standard-0.154 mMol of 1octanol in 10% ethanol (Merck KGaA, Darmstadt, Germany)-were carefully added. Just during the extraction time (approximately 1 h), the vial was heated up to 70 °C and stirred with a magnetic stirrer at 750 RPM, while the syringe was also kept at 70 °C. The extraction was accomplished by 15 strokes, with an aspired volume of 2000 µL per stroke, and flow rates of 10 µL/s up and 100 µL/s down. The 74 mm SPDE needle (Chromtech GmbH, Bad

HS-SPDE: Extraction of Volatiles in Infusions
The parameters for HS-SPDE were based on previous studies [21,22]. Within a 20 mL SPDE glass vial, 5 mL of the infusions and 100 µL of internal standard-0.154 mmol of 1-octanol in 10% ethanol (Merck KGaA, Darmstadt, Germany)-were carefully added. Just during the extraction time (approximately 1 h), the vial was heated up to 70 • C and stirred with a magnetic stirrer at 750 RPM, while the syringe was also kept at 70 • C. The extraction was accomplished by 15 strokes, with an aspired volume of 2000 µL per stroke, and flow rates of 10 µL/s up and 100 µL/s down. The 74 mm SPDE needle (Chromtech GmbH, Bad Camberg, Germany), coated with 50 µm of polydimethylsiloxane, activated carbon, and divinylbenzene (PDMS/AC/DVB, respectively) was coupled to a 2.5 mL syringe. After desorption in the GC port, the needle was flushed in the flush station with nitrogen gas at 270 • C for 15 min.

HS-SPDE: Extraction of Volatiles in the Mate Tea Samples
For identification, in order to maximize the extraction of volatiles present at low concentrations and generate enough of a MS signal, another approach had to be developed. To the best of our knowledge, no similar simple procedure is described in the literature to accomplish this task by HS-SPDE. A vial of 20 mL was filled with 1 mL of water at room temperature, closed just provisionally by pressing the cap against its top, and shaken for a few seconds to spread the water onto the internal walls. Then, 2 g of tea were added to the vial, the vial was sealed, and the tea was spread onto the internal walls while they were still humid by gently rotating and shaking the vial. The tea particles spread and adhered to the humid wall, creating a large surface area for the volatilization of compounds. All the other extraction parameters (i.e., regarding the strokes, the syringe, the needle, and the temperatures) were the same as listed above. These experiments were meant only for identification.

GC, FID, and MS Parameters
The gas chromatograph (GC) and flame ionization detector (FID) used for both analyses was a TRACE GC (Thermo Fisher Scientific GmbH, Dreieich, Germany) equipped with a Combi PAL autosampler (CTC Analytics AG, Zwingen, Switzerland) and operated under the parameters shown in Tables 1-3. The mass spectrometer (MS)-a TRACE DSQ (Thermo Fisher Scientific)-was coupled to the gas chromatograph described above and operated under the parameters shown in Table 4. Table 1. GC parameters for analyses of essential oils (from hydrodistillation).

Parameter Specification
Injected volume 1 µL Carrier gas Helium Carrier gas flow 1 mL/min (constant) Injection PTV, splitless Injection temperature 250 • C Temperature program 60 (Thermo Fisher Scientific)-was coupled to the gas chromatograph described operated under the parameters shown in Table 4.    Table 4.    (Thermo Fisher Scientific)-was coupled to the gas chromatograph descri operated under the parameters shown in Table 4. (Thermo Fisher Scientific)-was coupled to the gas chromat operated under the parameters shown in Table 4. (Thermo Fisher Scientific)-was coupled to the gas chromatograph described above and operated under the parameters shown in Table 4. (Thermo Fisher Scientific)-was coupled to the gas chromatograph descri operated under the parameters shown in Table 4. (Thermo Fisher Scientific)-was coupled to the gas chro operated under the parameters shown in Table 4.

Identification and Semi-Quantification of Compounds
Analyses were carried out using Xcalibur and Chrom Perfect software (Thermo Fisher Scientific/Axel Semrau, Sprockhoevel, Germany). The identification of the compounds was performed by a combination of a MS NIST library search and retention indices. The retention indices (RIs) were calculated by linear interpolation of the retention times (RTs) obtained for a sequence of n-alkanes (C8-C40 Alkanes Calibration Standard; Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany) analyzed under the same chromatographic conditions used for analyzing the essential oil, the HS-SPDE samples, and the analytical standards. The retention indices were mostly obtained from the compilation of Adams [23]. A minority of values not listed in this reference were obtained from the NIST database [24] for retention indices. Afterwards, the analytical standards were compared with their respective tentatively identified compounds in terms of their experimental retention indices and MS spectra.
The quantification was accomplished by peak area integration. The baselines were established as straight lines in between valleys where no specific masses but just background noise were detected. For chromatograms of essential oil, only the unidentified peaks with an average height (measured from the baseline) under 1000 mV were not considered for the calculation of the relative areas. For chromatograms of HS-SPDE, only peaks with a signal-to-noise ratio lower than 10 were not integrated but listed as 'trace'. Unidentifiable peaks had their peak area measured and included in the calculations of relative peak areas. Peaks identified as column or needle bleeding were completely disregarded.
Standard solutions containing 20 mg of standards and 10 mL of EtOH:H 2 O (1:9, v/v) were prepared and stored at 2 • C until analysis. The standards/chemicals are described in Table S1.

Replicates
For the essential oil samples, three chromatograms (three replicates) for each tea type were evaluated. Mean peak areas (n = 3) of the analytes were calculated. The calculation of relative areas comprised the unidentified compounds as well. For the HS-SPDE samples, the three chromatograms obtained from both tea types (A and B) in both infusion procedures (SI and TCI) were evaluated. For these samples, prior to the calculation of the relative areas, the areas of the analytes were normalized by dividing them by the area of the internal standard.

Compilation of Odor Thresholds of the Identified Volatiles
In order to achieve a better comprehension of the analytical data and assist in the identification of the potential odor active compounds, the odor thresholds of the identified compounds were compiled. The vast majority of the odor threshold values in water were obtained from the extensive compilations of van Gemert [25], and a few values from other separate references specified below the appropriate table of results. First, the minimum threshold values found in the literature were compiled separately, in order to compose the list of the 'minimum odor threshold values' ever reported in the literature. When several threshold values were available (which was the case for most of the compounds), minimum and maximum values were excluded and the final selection was based on the following criteria (when applicable): mode-values found repeatedly in the different references were preferred; and year of the reference-the most recent studies were preferred. No distinction was made (when mentioned) between detection thresholds and recognition thresholds. The aroma descriptors were mostly obtained from The Pherobase [26], when available, or from other references specified below the appropriate table of results.

Compounds Obtained by Hydrodistillation and SPDE
The compounds obtained by hydrodistillation (in the essential oils) and HS-SPDE (in the infusions) have their identification data shown in Table 5. Most of the peaks could be tentatively identified by a combination of a MS library search and retention indices. All these tentatively identified compounds matched the reference standards (which were tested) when comparing their mass spectra and retention time, being considered correctly identified. Some compounds that were identified with low MS library search matches and/or an imprecise RI match (in some cases, no RI data were found in the literature) are indicated with a question mark '(?)'. Regarding the unidentified compounds, the following is mentioned, instead of their names: either the most abundant masses (m/z) found in their spectra; or 'unknown', for peaks that did not display clearly distinguishable predominant masses. These peaks might be composed of mixtures of compounds, as reported by Purcaro et al. [5]. All the available data about each unidentified peak (m/z, RT, and RI) that may eventually be useful for future investigations are included.    [23], when just the numbers are mentioned; n-values from the NIST database [24]; b-values from the literature [2]; '(?)'-low MS library search matches and/or imprecise retention index correspondence; * values found for a similar column, other than DB-5MS; x-identification confirmed by comparison with the retention time and mass spectrum of the authentic standard; '-'-not detected or not available.
The relative peak areas of the volatiles detected by the different methods are represented in Figure 3, where only the major peaks are labeled. Complete information about all the semi-quantified compounds is provided in Table S2 (Supplementary Material).

Odor Thresholds
A rank of the average odor thresholds of the identified compounds is presented in Table 6, organized from the lowest to the highest average value, i.e., from the compounds that tend to be perceptible at lower concentrations to the ones that tend to be perceived just at higher concentrations, respectively. For some identified compounds, no information about threshold values nor odor characteristics was found. In this case, they were not included in this list.

Odor Thresholds
A rank of the average odor thresholds of the identified compounds is presented in Table 6, organized from the lowest to the highest average value, i.e., from the compounds that tend to be perceptible at lower concentrations to the ones that tend to be perceived just at higher concentrations, respectively. For some identified compounds, no information about threshold values nor odor characteristics was found. In this case, they were not included in this list.

Compounds Obtained by Hydrodistillation and SPDE
A remarkable difference between the extraction techniques employed for the present work can be observed at a first glance: HS-SPDE was more sensitive to the more volatile compounds and hydrodistillation to the less volatile. For instance, cymene is the 11th identified compound in the HS-SPDE samples while it is the fifth in the essential oils (Table 5). These results are logical, considering that HS-SPDE occurs at a lower temperature, and during a shorter extraction time, but within a hermetic vial, which prevents any losses of analytes. Some of the most volatile compounds in the essential oils could not be detected or identified either due to their low concentration or due to the saturated peak of the solvent (n-hexane) that covered them. Considering that mate tea traditional consecutive infusions are prepared while they are being drunk-therefore with warm but not boiling water [13]-hydrodistillation at 100 • C would be less representative than HS-SPDE at 70 • C, which was performed closer to the temperature of consumption. Independently of the temperature, HS-SPDE applied to the analysis of the infusions themselves presents the advantage of analyzing the final product that is ingested (the infusion) instead of the ingredient (mate tea) used for preparing the beverage.
In combination, the results from hydrodistillation and HS-SPDE are comparable with the findings of Bastos et al. [2]. This study presented 32 identified and semi-quantified volatiles, which creates a certain intersection between the results from the two different extraction techniques. It is important to emphasize that in this other study, the addition of a non-polar solvent to the distillation procedure, namely dichloromethane, might have assisted in preventing the loss of the most volatile components. These could be found at appreciable levels in the present study just in the HS-SPDE samples but not in the essential oils. Examples are: limonene, cymene, and eucalyptol, which comprised more than 30% of the area in the chromatograms from HS-SPDE, while in case of the essential oils this sum was lower than 0.5%.

Odor Thresholds
First of all, it must be noted that the odor threshold values compiled from the literature ( Table 6) are from different studies, which employed different methods and present variations of many folds. Furthermore, they were determined at room temperature using water as a matrix. In the case of mate tea chimarrão-type infusions, prepared with warm/hot water, these values change greatly once the vapor pressure of a substance increases exponentially with the temperature [11]. Therefore, caution is necessary when considering these data.
In between the values at the extremes of Table 6, a difference of 3.2 million fold can be observed. This indicates that the perception of the different volatiles is greatly driven by these values. Some of them possess such low odor thresholds that their occurrence above the threshold and consequent odor contribution already become likely upon their detection by a gas chromatograph, which is frequently less sensitive than the human olfactory system [11]. Examples of these compounds with extremely low values are: damascenone <(Z)-β->, damascone <(E)-β->, ionone <(E)-β->, 2-decenal <(E)->, octanal, ionene, <α->, hexanal, decanal <n>, ionone <(E)-α->, 2-octenal <(E)->, cyclocitral <β->, and linalool. On the other hand, compounds at the bottom of the table are more unlikely to impart their specific individual notes, e.g., camphene, methyl hexadecanoate, menthol, and menthone.
Bastos et al. [2] analyzed samples of Brazilian 'green mate' (supposedly also of the Brazilian chimarrão type) and 'chá-mate' (roasted), both from the same batch of raw materials and reported about volatile compounds in both samples. Some of the major compounds in the non-roasted samples are in accordance with those found in the present research and were lower or absent in the roasted samples, namely: pinene <α->; myrcene; limonene; linalool; terpineol <α>; geraniol; 2-decenal <(E)->; damascone <(E)-β->; and methyl hexadecanoate. These findings reinforce their presence in the list of typical major volatile compounds in this product.
In sum, a long list of compounds might be associated with the unique freshness of the Brazilian chimarrão mate tea. Remarkably, many terpenoids must be involved, even though several compounds from other classes of compounds are certainly inherent to its overall sensory profile. It is also important to consider that the freshness of chimarrão must be dependent not just on the presence of compounds imparting the typical fresher ('non-aged') notes at or above noticeable levels but also dependent on the concurrence of low levels or the absence of volatiles imparting the mature, aged, or roasted character. Many of these compounds (imparting aged notes) were described in other studies with other mate tea types [1,3,4,12]. Even though the results presented in this research are still inconclusive, they constitute a database to serve as a starting point for determining active and key odorants within the volatile fraction of this product in further future research, which should employ sensory analysis and other tools such as gas chromatography-olfactometry (GC-O).

Conclusions
In total, 85 compounds were identified (or tentatively identified) and semi-quantified in Brazilian chimarrão mate tea. Some compounds (mostly smaller peaks) remained unidentified. Approximately 50% of the identified compounds were commonly reported in studies with different mate tea types. Potential key odorants are supposed to be comprised within a list of numerous molecules (41) that seem to be specific to this product and are mostly composed of terpenoids (isoprenoids). The odor profile of this product (Brazilian chimarrão mate tea) must be characterized by: the presence of compounds imparting the typical freshness; and the absence or low levels of some compounds typically reported in other mate tea types, which derive from specific processes such as aging and roasting. Further investigations based on other tools such as GC-O and sensory analysis are necessary to define the key odorants in this product.

Supplementary Materials:
The following are available online at https://www.mdpi.com/article/10 .3390/separations8090131/s1, Table S1: Standards and chemicals used for GC-MS, Table S2: Relative peak areas of the compounds obtained by hydrodistillation and SPDE.