Characterization of Key Odor-Active Compounds in Sun-Dried Black Tea by Sensory and Instrumental-Directed Flavor Analysis

The aroma profile of sun-dried black tea (SBT) was identified by headspace solid–phase microextraction (HS–SPME) coupled with gas chromatography–mass spectrometry (GC–MS) and gas chromatography–olfactometry (GC–O). A total of 37 scents were captured by using the GC–O technique, and 35 scents with odor intensities ranging from 1.09 ± 1.93 to 9.91 ± 0.29 were identified. Twenty-one compounds were further identified as key odor-active compounds with odor activity values (OAVs) greater than or equal to one. These key odor-active compounds were restructured with their detected concentrations, and the aroma profile of the selected SBT sample was successfully imitated to a certain extent. An omission test was performed by designing 25 models and confirmed that (E)-β-damascenone, β-ionone, dihydro-β-ionone, linalool, and geraniol were the key odor-active compounds for the aroma profile of SBT. Meanwhile, phenylethyl alcohol, (E)-2-decenal, hexanal, and methyl salicylate were also important to the aroma profile of SBT. This study can provide theoretical support for the improvement of the aroma quality of sun-dried black tea.


Introduction
Tea has many kinds of health benefits. As a beverage, the consumption of tea ranks second only to water in the world [1]. Tea aroma is a prime criterion when evaluating tea quality, and it is also an important basis for consumers to choose. The formation of the tea aroma profile is the result of the comprehensive action of a variety of volatile compounds with different concentrations [2]. With the development and application of gas chromatography-mass spectrometry (GC-MS), more than 600 volatile compounds have been separated and identified in teas [3]. However, it is well known that not all volatile compounds have odors and contribute to the aromatic quality of finished product teas, and only a few of these compounds emit a scent and play a decisive role in the aroma profile of finished product teas, which are defined as the key odor-active compounds [4]. These odor-active compounds account for a small proportion of the complex volatile compounds, but they can dominate the overall aroma of teas. At present, how to effectively separate and identify these key odor-active compounds has become the focus of tea aroma research because the identification of key odor-active compounds in teas will have theoretical guidance for improving the aroma quality.
Sun-dried black tea (SBT), a kind of tea that evolved from Dianhong tea, is between Dianhong and Pu-erh tea, closer to Dianhong tea [5]. SBT is made from the leaves of In total 20 sun-dried black tea samples harvested in 2019 were collected from a tea market in Pu'er City (Pu'er, Yunnan, China) and Zhenyuan Taihe Sweet Tea Co., Ltd. (Pu'er, Yunnan, China), and 11 representative samples were selected by 6 experienced assessors according to the "Methodology for Sensory Evaluation of Tea (Chinese Standards, GB/T 23776-2018)" in the sensory analysis laboratory of Yunnan Tasly Deepure Biological Tea Group Co., Ltd. (Pu'er, Yunnan, China). All the representative samples were ground by a grinder with low temperature and passed through 30 mesh for future use. In order to improve the accuracy and applicability of this study, the experiment sample used in the optimization process was prepared by mixing an equal mass of 11 representative sun-dried black tea samples.

HS-SPME Procedures
The 65 µm polydimethylsiloxane/divinylbenzene (PDMS/DVB) fiber (Bellefonte, PA, USA) was selected and used for HS-SPME analysis. The extraction process was established as follows: sun-dried black tea (2.0 g), sodium chloride (2.0 g), water (6.0 mL), and a microstirring bar were put into a headspace vial. The vial was sealed and placed in a 50 • C (constant temperature) water bath. After 5 min equilibrium, the fiber was inserted into the vial and kept for 60 min at 120 rpm (stirring speed). After extraction, the fiber was immediately exposed to the GC injector (250 • C, 4 min) for desorption and analysis. Before the experiment five main factors (fiber, amount of water and sodium chloride, extraction temperature, and time) were optimized, and the data of the optimization process is supplied in the Supplementary Material (Figures S1-S5).

GC-O and GC-MS Analysis
An Agilent 7890A GC (Agilent Technologies, Palo Alto, CA, USA) equipped with an FID and a sniffing port (Sniffer 9000, Brechbühler, Switzerland) was used for GC-O analysis. An Agilent capillary column HP-5MS (30 m × 0.25 mm i.d., 0.25 µm film thickness; Agilent Technologies, Palo Alto, CA, USA) was used for the separation of target compounds. The conditions of the GC were set as follows: initial temperature 50 • C (3 min), increased to 180 • C at a rate of 3 • C min −1 and maintained for 5 min, and then increased to 250 • C at a rate of 10 • C min −1 , and maintained for 5 min; the temperature of the injection port, the injection mode and the flow velocity of carrier gas (constant-flow mode; He > 99.999%) were 250 • C, splitless, and 1.0 mL min −1 , respectively. The conditions of olfactometry were set as follows: the flow of humidified air was 60 mL min −1 and the temperature of the transfer line was 250 • C. A "Y" glass was used between the FID detector and the sniff port to split the GC effluent into 1:1.
In total, 6 assessors (4 males and 2 females, average age 32, non-smoking history, odor cosmetics were not used during the experiments) were selected from Yunnan Tasly Deepure Biological Tea Group Co., Ltd. (Pu'er, Yunnan, China) and Pu'er Institute of Pu-erh Tea among 17 experienced tea experts in accordance with the International Standard method [20] to perform GC-O analysis. Before the experiment, all selected assessors were trained for more than 20 h over a period of 30 d using 13 odor-active standards [α-terpineol, linalool, (E)-nerolidol, (3Z)-3-hexenyl 2-methylbutanoate, hexanal, 2-hexenal, methyl salicylate, phenylethyl alcohol, α-ionone, β-ionone, geraniol, (E)-β-damascenone and 1-octen-3-ol], which were detected in SBT, until all the assessors could quickly describe the aroma characteristics and aroma intensities of all selected odor compounds at their different diluted concentrations. The detection frequency method was used to analyze the aroma characteristics and aroma intensities of each effluent in GC-O analysis. A 9-point scale from 1 to 9, wherein "1", "5", and "9" indicated the weak odor, moderate odor, and extreme odor intensity, respectively, was used to quantify odor intensities of the effluents. During the experiments, the odor characteristics and odor intensities of the effluents were recorded and scored, respectively. Finally, the odor intensity value of each detected odor-active compound was averaged for all the scores that were detected by six assessors (three times by each assessor). After GC-O analysis, all the samples were analyzed by using GC-MS under the same capillary column, and the perceived scents were identified by the retention indices (calculated by using n-alkanes from C 8 to C 32 ) between the GC-O and the GC-MS, the mass spectra database (NIST11.L) of GC-MS, the odor characteristic of each odor-active compound, and corresponding standard of each odor-active compound.
The conditions of GC-MS were listed as follows: an Agilent 7890A GC equipped with 5975C mass selective detector quadrupole MS instrument (Agilent Technologies, Palo Alto, CA, USA) was used to identify the perceived odor-active compounds. The conditions of GC (injector temperature, oven temperature program, the kind of carrier gas, and the flow rate of carrier gas) and the model capillary column were the same as described under GC-O analysis. The electron-impact (EI), interface temperature, ion source temperature, quadrupole temperature, mass scan range, and solvent delay time were 70 eV, 280 • C, 230 • C, 150 • C, 35-450 aum, and 3.0 min, respectively.

Quantitation of Odor-Active Compounds and OVA Calculation
Considering that non-volatile components may affect the aroma emission, a deodorized tea (odor-blank matrix) was prepared according to the method of the previous report [21] with minor modification during the quantitative process of odor-active compounds. The preparation process was as follows: 100 g SBT and 800 mL ultrapure water were weighed into a 2000 mL rotary evaporation bottle, following which the tea odor-active compounds were rotary evaporated in a 60 • C water bath. The process was repeated many times (greater than or equal to 5 times) until the sensory flavor of the tea was very light. Thereafter, the SBT was dried using an oven at 60 • C. The deodorized SBT, internal standard (n-decanol), and corresponding standards of odor-active compounds were used to obtain the quantitative result of each target compound. The process was listed as follows: initially, the authentic standards were dissolved using anhydrous alcohol and then diluted with ultrapure water to different concentrations. The dissolved standards dilution, internal standard (0.1 mL n-decanol, 4.15 µg/mL), and deodorized tea (2.0 g) were placed into a 20 mL headspace vial (Agilent Technologies, Palo Alto, CA, USA) and sealed. The vial was kept for 24 h at 4 • C and prepared with the HS-SPME method, as mentioned above (Section 2.3.). The quantitative standard curves of target compounds were established by plotting the peak area ratio against the corresponding concentration ratio.
Odor activity value (OAV) is often used to evaluate the contributions of odor-active compounds and further screen the odor-active compounds that play a decisive role in the aroma profile of the analyzed samples [22]. The OAVs of compounds are equal to the ratio of corresponding concentrations to their respective odor thresholds in water, and compounds with an OAV greater than or equal to one are generally considered to contribute significantly to the aroma profile of the analyzed samples [23]. In this study, the odor thresholds of detected odor-active compounds were obtained from literature or books (References [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17], supplied in Supplementary Material).

Aroma Recombination and Omission Test
For the aroma recombination experiment, the deodorized tea (treatment process is consistent with 2.5) was used as a blank matrix. The odor-active compounds with OAVs of greater than or equal to one that detected in a randomly selected SBT were added into the deodorized tea according to their detected concentrations and mixed well. The recombination sample was sealed and kept for 24 h at 4 • C before sensory analysis.
The contributions of a single odor-active compound and a group of odor-active compounds were evaluated using an omission test. Twenty-five aroma models were prepared, and the aroma characteristic and aroma intensities of the reduced models were evaluated against two complete recombination samples with a triangle test according to the method of a reported study [24]. All the recombination samples were coded with random three-digit numbers, and all the assessors were asked to select different ones through sensory analysis.

Descriptive Sensory Analysis
A total of 10 experienced assessors (6 males and 4 females, average age 30, with non-smoking history, odor cosmetics were not used during the experiments) were selected from Yunnan Tasly Deepure Biological Tea Group Co., Ltd. (Pu'er, Yunnan, China) and Pu'er Institute of Pu-erh Tea based on the International Standard method [20] to perform the descriptive sensory analysis of SBT and recombination samples. The procedure of descriptive sensory analysis referenced the method described by the previous study with some modifications [7]. In total, 6 odor terms (floral, fruity, green/grass, sweet, woody, and unpleasant flavor) of SBT and corresponding authentic reference standard (detected in SBT; linalool for a floral note, (Z)-3-hexenyl isovalerate for a fruity note, hexanal for green/grass odors, n-hexyl caproate for a sweet note, β-ionone for a woody note) were determined by 20 assessors through 3 preliminary sessions before the sensory analysis. A 9-point scale from 1 to 9 was used, wherein "1", "5", and "9" indicated the weak odor, moderate odor, and extreme odor intensity, respectively.
The descriptive sensory analysis was performed according to the "Methodology for Sensory Evaluation of Tea (Chinese Standards, GB/T 23776-2018)". SBT (3.0 g) was placed in a special teapot and infused with 150 mL of freshly boiled water (kept for 5 min before sensory analysis). All the samples were coded with three-digit numbers and randomly provided to the assessors after brewing. The intensity value of each odor term was the average scores from 10 assessors (3 times by each assessor).

Statistical Analysis
Data analysis and figure drawing were performed by using MS Excel 2010 and Origin-Pro software (version 9.5.1, OriginLab Inc., Northampton, MA, USA), respectively.

GC-MS Results for Sun-Dried Black Tea
Prior to GC-O analysis, the volatile compounds in 11 representative SBT samples were analyzed by HS-SPME/GC-MS, and the representative total ion chromatography of SBT was presented in Figure 1. A total of 86 volatile compounds were tentatively identified according to their mass spectrum and RI, including 21 hydrocarbons, 17 alcohols, 14 aldehydes, 13 ketones, 10 esters, 3 organic acids, 2 phenols, 2 ethers, 1 lactone, 1 furan, 1 nitrogen compound, and 1 oxygen heterocyclic compound, which were listed in Table  S1. It was reported that 16 hydrocarbons, 17 alcohols, 11 aldehydes, 13 ketones, 10 esters, 3 organic acids, 1 ether, 1 lactone, 2 furans, and 2 nitrogen compounds were detected in a previous study [8]. The main volatile compounds in our study were consistent with those reported by Lv et al. [8].  to "×

Odor-Active Compounds Determination Using GC-O
In order to gain deep insights into the odor characteristics of the volatile compounds and further select the key odor-active compounds from the volatile compounds, GC-O analysis was performed. A total of 37 scents were perceived by GC-O in 11 SBT samples, and 35 compounds were identified using GC-MS based on their mass spectrum, RIs, odor characteristics, and authentic standards, and 2 compounds with moderate odor were not identified (Table 1). Among these detected compounds, β-ionone (AI: 9.91 ± 0.29), (E)-β-damascenone (AI: 9.64 ± 0.48), linalool (AI: 8.64 ± 0.98), linalool oxide II (AI: 8.45 ± 0.78), epoxylinalol (AI: 7.91 ± 0.79), linalool oxide I (AI: 7.82 ± 0.57), dihydroβ-ionone (AI: 7.55 ± 0.99), geraniol (AI: 7.36 ± 1.55), methyl salicylate (AI: 6.82 ± 1.11), α-ionone (AI: 6.36 ± 0.98), phenylethyl alcohol (AI: 5.64 ± 0.77), geranyl acetone (AI: 5.64 ± 0.97), and (E)-nerolidol (AI: 5.45 ± 1.87) had higher aroma intensities, which indicated that these compounds played an important role in the formation of SBT aroma. α-Ionone, β-ionone and dihydro-β-ionone, a group of carotenoid-derived aroma compounds, all presented "woody" and "violet-like" scents, which were detected in sun-dried Pu-erh tea of ancient tea plants from Bulang Mountain [25]. (E)-β-Damascenone presented "rose-like" and "sweet" scents, with low human odor perception threshold, it was reported that (E)-β-damascenone had a great contribution to the aroma profile of black teas [26]. Linalool presented pleasant "floral" and "citrus-like" scents, which were not only detected as key odor-active compounds in green tea, white tea, yellow tea, oolong tea, black tea, and dark tea samples [27] but also detected in flowers and spice plants [28,29]. Linalool oxides (linalool oxide I, linalool oxide II and epoxylinalol) presented "floral" and "woody" scents, which were detected with a high concentration in semi-fermented and fermented teas [27]. Geraniol, phenylethyl alcohol, and (E)-nerolidol all presented a "rose-like" scent, which was detected in many kinds of tea samples. Geraniol and (E)-nerolidol were reported as the key odor-active compounds for the quality of black teas [30], and (E)-nerolidol was also considered to be one of the key odor-active compounds for the high-quality of oolong tea [31]. Geranyl acetone presented "rose-like" and "green" scents, which were detected in black and ripened Pu-erh teas, and were considered to play an important role in the aroma profile of black and ripened Pu-erh teas [32,33]. Methyl salicylate presented "mint" and "wintergreen-like" scents, which were detected only in the teas that had a least a medium-degree fermentation but could not be detected in the green and lightly fermented teas, and it was reported that methyl salicylate and (E)-2-hexenal together could be used as an index to differentiate semi-and fully-fermented teas [34].

Quantitation of Odor-Active Compounds and OVA Calculation
The odor-active compounds from the bulk of volatile compounds of SBT were identified by using a unique technique of odor-active compounds screening. To deeper evaluate the aroma contributions of these perceived odor-active compounds, their contents were quantitated in 11 SBT samples based on internal standard and their corresponding standards, and the quantitative results of these odor-active compounds supplied in Table 2 and  (25.56 ± 52.14 µg/kg), geranic acid (25.13 ± 38.39 µg/kg), dihydro-β-ionone (24.86 ± 7.32 µg/kg), and 1-octen-3-ol (23.52 ± 49.27 µg/kg) in SBT samples were relatively low. These odor-active compounds were the first quantified in SBT. The contributions of odor-active compounds to the aroma profile of SBT not only depend on its content but also depends on its odor threshold value. Odor-active compounds with OAVs greater than or equal to one were usually considered to play a major role in the aroma profile of the samples [35]. To further deeply elucidate the contribution of the odor-active compounds in SBT, their OAVs were calculated using the quantitative results and reported thresholds, and the calculation results are shown in Tables  greater than 1, and indicated that these compounds played an important role in the aroma profile of SBT. Among them, alcohols, ketones, hydrocarbons, and aldehydes occupied most of the components, and the results of this study were consistent with those previously reported [8]. However, the OAVs calculation showed that an odor-active compound with higher content did not mean a larger OAV or vice versa. For example, dihydro-βionone (24.86 ± 7.32 µg/kg) was present at low concentrations in SBT samples, but its OAV was 24856.37 (its threshold was 0.001 µg/kg in water). By contrast, epoxylinalol (973.71 ± 371.22 µg/kg) was present at a relatively high concentration in SBT samples, but its OAV was only 0.32. In general, the results of OAVs and GC-O were basically consistent with each other.

Aroma Recombination and Omission Test
To confirm the contributions of these compounds with OAVs greater than or equal to one, the aroma recombination test was designed and performed. The deodorized SBT was used as a blank matrix. An SBT sample was randomly selected as a reference for aroma recombination. Then, 21 compounds with OAVs greater than 1 were added to the deodorized tea based on the quantitative results of the reference sun-dried black tea and mixed well. The recombination sample was sealed and kept for 24 h at 4 • C before sensory analysis. The recombination sample and selected reference sun-dried black tea were evaluated by 10 experienced assessors according to the "Methodology for Sensory Evaluation of Tea (Chinese Standards, GB/T 23776-2018)" described in Section 2.7. Six odor terms, including floral, fruity, green/grass, sweet, woody, and unpleasant flavor were identified as the main aroma characteristics based on the sensory results of SBT samples. The sensory results of the recombination sample and selected reference sample are shown in Figure 2. The results showed that the odor profile of the SBT could be successfully imitated to a certain extent by combining these 21 odor-active compounds at their detected concentrations in the selected reference sample, specifically in terms of its floral, fruity, green/grass, and sweet scents. The recombination sample was more inclined to a woody scent. This phenomenon may be due to the conversion of the non-volatile components while preparing the deodorized matrix or the difference in the extraction amounts of odoractive compounds while brewing (the odor-active compounds of the recombination sample were easier to give off). Overall, the results of odor recombination further verified the accuracy of the identified key odor-active compounds in SBT.
In order to deeply verify the contributions of a single odor-active compound with OAVs greater than or equal to 1 to the aroma profile of SBT, the omission test was performed by designing 25 models, in which a single odor-active compound or a group of odor-active compounds was omitted. Taking the recombination sample as a reference, 25 omission models were evaluated by triangle test, and the results are supplied in Table 3.
Model 1 showed that 10 assessors were able to perceive the entire omission of all alcohols (including linalool, geraniol, phenylethyl alcohol, (E)-nerolidol, linalool oxide II, 1-octen-3-ol, linalool oxide I, and α-terpineol) with a very high significance (p ≤ 0.001), indicating that these alcohols played a major role in the aroma formation of SBT. Compared with the recombination sample, the single omission of linalool (model 1-1), geraniol (model 1-2), or phenylethyl alcohol (model 1-3) showed a very high significant difference (p ≤ 0.001), a high significant difference (p ≤ 0.01), and significant difference (p ≤ 0.05), respectively. While the single omission of (E)-nerolidol (model 1-4), linalool oxide II (model 1-5), 1-octen-3-ol (model 1-6), linalool oxide I (model 1-7), or α-terpineol (model [1][2][3][4][5][6][7][8] showed no significant difference (p > 0.05), indicating that the effect of these compounds on the aroma profile of SBT is relatively small. tions in the selected reference sample, specifically in terms of its floral, fruity, green/grass, and sweet scents. The recombination sample was more inclined to a woody scent. This phenomenon may be due to the conversion of the non-volatile components while preparing the deodorized matrix or the difference in the extraction amounts of odor-active compounds while brewing (the odor-active compounds of the recombination sample were easier to give off). Overall, the results of odor recombination further verified the accuracy of the identified key odor-active compounds in SBT. In order to deeply verify the contributions of a single odor-active compound with OAVs greater than or equal to 1 to the aroma profile of SBT, the omission test was performed by designing 25 models, in which a single odor-active compound or a group of odor-active compounds was omitted. Taking the recombination sample as a reference, 25 omission models were evaluated by triangle test, and the results are supplied in Table 3. Table 3. The results of the omission test.
Model 4 showed that only 5 assessors were able to detect the entire omission of all hydrocarbons (1-methyl-naphthalene and D-limonene) with no significant difference compared with the recombination sample. Model 5 showed that 7 assessors were able to detect the single omission of methyl salicylate with a significant difference (p ≤ 0.05). Model 6 showed that only 3 assessors were able to detect the single omission of 2-pentyl-furan with no significant difference (p > 0.05).

Conclusions
The odor profile of sun-dried black tea was successfully identified by sensory and instrumental-directed flavor analysis. A total of 37 scents were perceived under GC-O analysis, and 35 compounds with odor intensities ranging from 1.09 ± 1.93 to 9.91 ± 0.29 were identified. Twenty-one of them were further identified as key odor-active compounds with OAVs greater than or equal to one based on their quantitative results and thresholds. These key odor-active compounds were restructured with their detected concentrations in a reference sample, and the aroma profile of the selected sun-dried black tea sample was successfully imitated to a certain extent. An omission test was performed by designing 25 models and confirmed that linalool, dihydro-β-ionone, β-ionone, and geraniol were the key odor-active compounds for the aroma profile of sun-dried black tea. Meanwhile, phenylethyl alcohol, (E)-β-damascenone, (E)-2-decenal, hexanal, and methyl salicylate were also important to the aroma profile of sun-dried black tea.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/foods11121740/s1, Figure S1: The extraction efficiency of four fibers; Figure S2: Influence of water amount on the number and sum of intensity value of aroma-active compounds; Figure S3: Influence of NaCl amount on the number and sum of intensity value of aroma-active compounds; Figure S4: Influence of extraction temperature on the number and sum of intensity value of aroma-active compounds; Figure S5: Influence of extraction time on the number and sum of intensity value of aroma-active compounds; Table S1: Volatile compounds detected by GC-MS in sun-dried black tea; Table S2: The quantitative results of odor-active compounds; Table S3: The OAVs of odor-active compounds.