Investigation of the Effect of Fragrance-Enhancing Temperature on the Taste and Aroma of Black Tea from the Cultivar Camellia sinensis (L.) O. Kuntze cv. Huangjinya Using Metabolomics and Sensory Histology Techniques

Abstract

Huangjinya has recently seen widespread adoption in key tea-producing areas of China, celebrated for its unique varietal traits. Its leaves are also used to produce black tea with distinctive sensory characteristics. The fragrance-enhancing (EF) process is essential in crafting Huangjinya black tea (HJYBT) and is significant in flavor development. However, the impact of EF on non-volatile metabolites (NVMs), volatile metabolites (VMs), and their interactions remains poorly understood. This study aims to investigate how EF temperatures (60 °C, 70 °C, 80 °C, 90 °C, and 110 °C) influence HJYBT flavor transformation. Quantitative descriptive analysis revealed that EF improved the color, aroma, and appearance of tea leaves. Moreover, after an EF temperature of 80 °C, the HJYBT exhibited lower bitterness and astringency, whereas floral, sweet, and fruity aromas became stronger. However, when EF temperatures exceeded 90 °C, a pronounced burnt aroma developed, with HJYBT at 100 °C exhibiting caramel and roasted notes. Partial least squares discriminant analysis indicated that geraniol and linalool contribute to floral and fruity aromas, while 2-ethyl-6-methyl-pyrazine, furfural, and myrcene are key volatiles for caramel and roast aromas. Heptanal, methyl salicylate, α-citral, 1-hexanol, and (E)-3-hexen-1-ol were found to modify the green and grassy odor. Overall, HJYBT treated at 80 °C EF exhibited the highest umami, sweetness, floral and fruity aromas, and overall taste, exhibiting the least astringency, bitterness, and green and grassy notes. These results provide a significant theoretical basis for enhancing HJYBT quality and selecting the optimal EF method.

1. Introduction

Black tea is noted for its bright red liquor, sweet and robust aroma, and rich flavor, as well as its anti-inflammatory and anti-obesity activity, cardiovascular disease prevention, and antidiabetic effects [1,2,3,4,5]. Consequently, black tea is favored worldwide, accounting for approximately 75% of global tea consumption [3,6]. The processing technology of black tea includes withering, rolling, fermentation, drying, and fragrance enhancement (EF). Withering and EF play a significant role in determining the overall quality of black tea [2,6,7]. Recent research has focused on withering’s influence on black tea quality, exploring various methods and their effects on high-quality production [8,9,10]. EF technology, traditionally used in oolong tea processing, especially rock tea, exerts a decisive influence on developing characteristic flavors like the floral aroma of rock tea. EF promotes the formation of pyrrole, pyrazine, and benzaldehyde, which produce roast and caramel aromas in rock tea, thus enhancing its quality [11,12]. Charcoal roasting EF markedly intensifies the floral and fruity aromas of Lu’an Guapian green tea [13]. Additionally, EF aids in the production of roasting aroma compounds such as 2,5-dimethylpyrazine, ethylpyrazine, and trimethylpyrazine in Huangda tea [14,15]. However, with the continuous expansion of the black tea market, new EF technologies have emerged, and this has also led to variations in product quality [6]. Therefore, it is essential to thoroughly investigate the effects of EF on black tea quality to maintain consistent excellence.

EF entails reducing tea leaves’ moisture content to around 7% by mass, then cooling them and allowing them to regain moisture before subjecting them to controlled roasting. This procedure induces moisture loss and triggers intricate thermochemical reactions, such as the Maillard reaction, thermal degradation, deamination, and cyclization. These reactions modify the biochemical components’ structure and produce new compounds, which elevate the tea’s aroma and augment its quality [3,16]. Temperature exerts a profound influence on the EF process by modulating thermochemical reaction pathways, leading to the production of various volatile and non-volatile compounds, which shape the flavor profile of black tea [3]. For example, EF at 90 °C enhances the mellowness and sweetness of Yunnan black tea’s aroma [17], whereas EF at 70–75 °C imparts a sweet, mellow taste and aroma to Jinjunmei black tea [18]. Higher EF temperatures accelerate the Maillard reaction, enriching the tea leaves’ aroma with roasted, sweet, and caramel notes. Excessively high EF temperatures generate burnt odors, leading to a decline in tea quality [19]. Additionally, elevated EF temperatures decrease amino acids and theaflavins while increasing ester catechins, negatively impacting the quality of black tea [20]. Conducting a comprehensive metabolite analysis is vital to assess how varying EF temperatures influence black tea quality, ensuring consistency. Moreover, existing research rarely explores the impact of EF on the quality of mutant tea leaves, such as purple and albino varieties.

Huangjinya, a photosensitive mutant albino tea variety with golden yellow shoots, transitions from yellow to green with increasing light intensity and leaf maturity [21,22]. Extensively cultivated in China, Huangjinya has elevated amino acid levels and reduced tea polyphenols and catechins compared to conventional green tea varieties. As a result, green tea derived from Huangjinya is distinguished by its heightened freshness and diminished bitterness, conferring substantial economic value [22,23]. In addition, compared with general green tea cultivars, Huangjinya black tea (HJYBT) has a more umami and mellow taste, orange-red color, and sweet or floral aroma [24]. Advanced targeted metabolomics studies, using UPLC coupled with high-resolution mass spectrometry, have clarified the metabolite alterations during the processing of Huangjinya black tea (HJYBT) [21]. Moreover, HJYBT extract demonstrates preventive effects against obesity, metabolic disorders, neurodegenerative diseases, and modulates intestinal flora [24,25]. However, the specific impact of EF on HJYBT quality, especially concerning the key volatiles related to sensory quality, remains unresolved.

Quality characteristics of black tea encompass appearance, soup color, aroma, taste, and leaf bottom. Aroma and taste, which constitute 55% of the overall sensory quality, are predominantly determined by volatile and non-volatile metabolites [2,3]. Non-volatile compounds, such as catechins, flavonoids, caffeine, free amino acids, theaflavins, and thearubigins, are key indicators of tea quality [26]. The ratio of tea polyphenols to amino acids directly correlates with the umami and bitterness profiles of tea [27]. Theaflavins, thearubigins, and caffeine are key determinants of the “cream down” phenomenon in black tea. High-quality black tea exhibits elevated levels of these compounds, maintaining a typical theaflavins-to-thearubigins ratio of approximately 1:10 [28]. Tea processing involves the release or degradation of volatile compounds, profoundly influencing sensory characteristics and overall product quality, thereby affecting consumer preferences and acceptance [2]. The EF process modifies the composition and concentration of volatile compounds, resulting in the convergence and aggregation of essential aroma components, which substantially impacts black tea aroma. Sensory evaluation, a conventional method for tea quality assessment, relies on the expertise of evaluators. However, this method is subjective, time-consuming, and lacks scientific precision [29]. Quantitative description analysis (QDA) is a kind of descriptive sensory analysis method that is applied in research and development and for the test analysis of food flavor. QDA provides analysis, description and quantitative test functions, which can describe the sensory impression of a sample in the simple and common vocabulary. It has been widely used in the descriptors of the flavors of black tea [2,30].

Recent advancements in metabolomics have profoundly deepened the comprehension of plant chemical constituents and food flavor chemistry. This method provides thorough characterization and quantification of small molecule metabolites in tea, playing a key role in research on tea quality and processing [2,31,32]. High-performance liquid chromatography (HPLC), with its superior sensitivity and efficiency, allows for precise quantitative analysis of essential compounds such as catechins and amino acids in tea [13]. Gas chromatography–time of flight mass spectrometry (GC-TOFMS) provides exceptional sensitivity, resolution, peak capacity, and qualitative accuracy, making it indispensable for volatile metabolite analysis in tea [10,33]. Simultaneously, the electronic nose (E-nose), with its multiple sensor arrays, detects aroma components, facilitating the identification of aroma changes during tea processing and distinguishing tea aroma quality [29]. To investigate the impact of EF on HJYBT quality, EF was applied at various temperatures (60 °C, 70 °C, 80 °C, 90 °C, and 100 °C) to crude black tea processed through standardized withering, rolling, fermentation, and drying techniques. Quantitative analysis of flavor substances was conducted using spectrophotometry and HPLC, while volatile components at different EF temperatures were analyzed with E-nose and GC-TOFMS. Multivariate statistical analysis assessed the effect of EF temperature on both volatile and non-volatile metabolites and explored potential correlations between sensory quality characteristics and metabolites. The objective was to understand how different EF temperatures influence HJYBT flavor quality and to provide a theoretical basis for optimizing the EF process to enhance black tea quality.

2. Materials and Methods

2.1. Manufacturing of Black Tea Samples

All HJYBT samples were produced at the Sichuan Congou Engineering and Technology Research Center, Yibin Vocational and Technical College. Fresh one-bud and two-leaf Huangjinya leaves, sourced from Yibin Jiang’an Zhudu Tea Co., Ltd. (Yibin, China), underwent a five-step process: (1) Withering: Forty kilograms of fresh leaves were evenly spread on the withering trough, subjected to 26–28 °C and 70–75% relative humidity, and turned every 4.5 h until the moisture content reached 60–64%, a process taking approximately 22 h. (2) Rolling: The withered leaves were processed in a rolling machine for 80 min, comprising 20 min without pressure, 20 min of light pressure, 15 min of medium pressure, 15 min of heavy pressure, and 10 min of light pressure. (3) Fermentation: Rolled tea leaves were placed in a black tea fermentation box for 4.5 h at 25–28 °C and ≥95% relative humidity, with a leaf pile thickness of 15 cm. (4) Drying: The fermented leaves were initially dried at 115 °C for 15 min using a hot air dryer, reaching a moisture content of 20–25%. After cooling for 60 min, a second drying at 75 °C with a tea aroma machine reduced moisture content to 5%, producing raw tea. (5) EF treatment involved aromatizing the raw tea at 60 °C (T60), 70 °C (T70), 80 °C (T80), 90 °C (T90), and 100 °C (T100) for 3 h using the same tea aroma machine, three replicates were performed for each steps. Equipment was sourced from Zhejiang Shangyang Machinery Co., Ltd. (Quzhou, China). Three replicates of tea samples were collected at each EF temperature, with portions allocated for sensory evaluation and stored at −20 °C for metabolomics and E-nose analysis.

2.2. Standard Compounds and Chemicals

Standards for (+)-catechin (C), (−)-epicatechin (EC), (−)-gallocatechin (GC), (−)-epigallocatechin (EGC), (−)-gallocatechin gallate (GCG), (−)-epigallocatechin gallate (EGCG) and (−)-epicatechin gallate (ECG), (−)-catechin gallate (CG), gallic acid (GA), caffeine (CAF), myricetin (Myr), quercetin (Que), luteolin (Lut), kaempferol (Kae) and rutin (Rut), asparagic acid (Asp), glutamic acid (Glu), L-asparagine (L-asp), serine (Ser), glutamine (Gln), histidine (His), Glycine (Gly), threonine (Thr), arginine (Arg), alanine (Ala), theanine (Thea), tyrosine (Tyr), cystine (Cys), valine (Val), methionine (Met), tryptophan (Try), isoleucine (Iso), phenylalanine (Phe), lysine (Lys), and leucine (Leu) were purchased from Chendu Must Biotechnology Co., Ltd. (Chengdu, China). HPLC-grade methanol (MeOH) and acetonitrile (ACN) were purchased from Sigma-Aldrich (St. Louis, MO, USA). OPA regent (10 mg/mL) and borate buffer (0.4 M, pH 10.4) used for derivatization were from Agilent (Agilent Technologies, Palo Alto, CA, USA). All the reference compounds and reagents for untargeted metabolomics analysis were provided by Biomarker Technologies Co., Ltd. (Beijing, China). Other related chemicals were of analytical grade. Ultrapure water was used throughout the experiment.

2.3. Quantitative Descriptive Analysis

QDA assessed the sensory characteristics of all HJYBT samples [2,21]. An evaluation team of 9 certified reviewers (3 males and 6 females, aged 24 to 40 years), each holding a national senior tea appraiser certification, conducted the assessment according to the GB/T 23776-2018 [34]. The procedure involved placing 3 g of tea into an evaluation cup, adding 150 mL of boiling water, covering, and steeping for 5 min. The tea infusion was filtered into an evaluation bowl, and samples were randomly coded with 3-digit numbers before being distributed to panel members. Panelists independently recorded the flavor and aroma intensity of the tea in a clean, odorless environment maintained at 24–26 °C. The aroma and taste characteristics of HJYBT were assessed, with descriptions using terms such as green odors, floral, fruity, sweet, honey, caramel, and cooked aroma. Taste characteristics included sweetness, umami, bitterness, sourness, mellowness, and astringency [2,3,21].

2.4. Analysis of Chemical Compounds in Tea Leaves by Spectrophotometry and HPLC

Each HJYBT sample at different temperatures of EF was well dispersed and milled into fine powder (40 mesh). The contents of water extractions (WEs), tea polyphenols (TPs), free amino acids (FAAs), soluble sugar (SS), theaflavins (TFs), thearubigins (TRs), and theabrownins (TBs) in HJYBT were analyzed by a spectrophotometric method, as described previously [30,35]. The catechins, flavones, gallic acid, and caffeine contents were extracted as described by Liu et al. [1] and then quantitatively analyzed by a Shimadzu LC-16 HPLC system (Shimadzu, Japan) equipped with a photodiode array (PDA) detector. Chromatographic separation was performed with an Elite Hypersil BDS C18 column (150 mm × 4.6 mm, 3 μm). The elution conditions were maintained as follows: oven temperature: 43 °C; injection volume: 10 μL; flow rate: 1.0 mL/min; mobile phase A: 0.26% phosphoric acid, 5% ACN, and 94.74% H₂O; mobile phase B: 0.26% phosphoric acid, 80% MeOH, and 17.4% H₂O; detection wavelength: 278 nm and 370 nm. The gradient elution for the system was as follows: 0–3 min, 5%B; 3–22 min, 32%B; 22–26 min, 32%B; 26–29 min, 5%B and 29–40 min, 5%B.

2.5. Electronic Nose (E-Nose) Data Collection and Analysis

The volatile compounds of HJYBT were analyzed using a portable electronic nose system, PEN3 (Airsense Analytics GmbH, Schwerin, Germany), equipped with 10 different aroma receptors, including W1C, W5S, W3C, W6S, W5C, W1S, W1 W, W2S, W2 W, and W3S, which are sensitive toward the functional groups of “aromatic hydrocarbon”, “broad range (especially nitrogen oxide compounds)”, “aromatic ammonia”, “hydrogen”, “arom-aliph”, “broadmethane”, “sulfur-organic”, “broad-alcohol”, “sulfur-chlor”, and “methane-aliph”, respectively [36,37]. Before detection, 10 g of HJYBT sample was weighed and placed in a 100 mL conical flask, and 50 mL of boiling water was added and sealed with bilayer cling film immediately. The conical flask was left undisturbed for 10 min to ensure the equilibration of gases within the container. Then, zero gas (indoor air filtered with standard activated carbon) was pumped into the cleaning channel and sensors were reset. Sampling was carried out with a measurement time of 60 s and a pre-sampling time of 5 s, and the injection flow rate was set to 240 mL/min [29]. Each sample was analyzed six times.

2.6. Extraction and Identification of Volatile Compounds in HJYBT

2.6.1. Extraction of Volatiles Using Solvent-Assisted Flavor Evaporation

Place 30 ± 1 mg of HJYBT powder into 2 mL EP tubes. Extract with 0.24 mL of extraction liquid (Vmethanol: VdH₂O = 3:1), adding 10 μL of adonitol (1 mg/mL stock in dH2O) as the internal standard. Vortex for 30 s, then homogenize in a ball mill for 4 min at 45 Hz, followed by ultrasound treatment for 5 min while incubating in ice water. Centrifuge at 13,000 rpm and 4 °C for 15 min. Transfer 0.18 mL of the supernatant into a fresh 2 mL GC/MS glass vial. Take 25 μL from each sample to pool as a quality control (QC) sample, then dry completely in a vacuum concentrator without heating. Add 30 μL of methoxyamine hydrochloride (20 mg/mL in pyridine) and incubate for 30 min at 80 °C. Add 50 μL of the BSTFA reagent (1% TMCS, v/v) to the sample aliquots and incubate for 1.5 h at 70 °C. When cooled to room temperature, add 5 μL of FAMEs (a standard mixture of fatty acid methyl esters, C8–C16: 1 mg/mL; C18–C24: 0.5 mg/mL in chloroform) to the QC sample. Analyze all samples by GC-TOFMS.

2.6.2. GC-TOFMS Analysis

GC-TOFMS analysis utilized an Agilent 7890 gas chromatograph system coupled with a Pegasus HT time-of-flight mass spectrometer. The system employed a DB-5MS capillary column with 5% diphenyl cross-linked with 95% dimethylpolysiloxane (30 m × 250 μm inner diameter, 0.25 μm film thickness; J&W Scientific, Folsom, CA, USA). A 1 μL aliquot of the analyte was injected in splitless mode. Helium served as the carrier gas, with a front inlet purge flow of 3 mL/min and a column gas flow rate of 1 mL/min. Initially, temperature was controlled at 50 °C for 1 min, increased to 310 °C at a rate of 10 °C/min, and held at 310 °C for 8 min. The injection, transfer line, and ion source temperatures were set at 280 °C, 270 °C, and 220 °C, respectively. The electron impact mode operated at −70 eV. Mass spectrometry data were acquired in full-scan mode with an m/z range of 50–500 at a rate of 20 spectra per second following a solvent delay of 6.12 min.

2.6.3. Data Preprocessing and Analysis

Chroma TOF 4.3X software (LECO Corporation, St. Joseph, MI, USA) and the LECO-Fiehn Rtx5 database were utilized for raw peak extraction, baseline filtering and calibration, peak alignment, deconvolution analysis, peak identification, and peak area integration. Metabolite identification considered both mass spectrum and retention index matches. Peaks detected in fewer than 50% of QC samples or with an RSD greater than 30% in QC samples were removed. Following normalization of the original peak area data to the total peak area, further analysis was conducted.

2.6.4. Quantification and Odor Activity Values Calculation

A semi-quantitative method using an internal standard (adonitol) was employed for the relative quantification of volatile compounds following the formula established in a previous study [2]. Odor activity value (OAV) > 1 is generally considered necessary to ascertain the contribution of volatile compounds to aroma. OAVs were calculated by the ratio of each compound’s relative concentration to its odor threshold in water [2,33].

2.7. Data Statistics and Analysis

One-way analysis of variance (ANOVA) with least significant difference (LSD) and correlation analysis were performed by SPSS (version 22.0; Chicago, IL, USA). Principal component analysis (PCA), hierarchical cluster analysis (HCA), and orthonormal partial least-squares discriminant analysis (OPLS-DA) were performed using SIMCA (version 14.1; Umetrics, Umea, Sweden). The p-values were ANOVA results, and variable important in projection (VIP) values were extracted from the OPLS-DA result. Significantly different metabolites (DVCs) were determined by VIP > 1 and p < 0.05. Images of the data were visualized using GraphPad Prism (version 9) and integrative toolkit-TBtools. Flavor annotations for DVCs use the TGSC website (http://www.thegoodscentscompany.com/ (accessed on 1 September 2024)) and FEMA website (https://www.femaflavor.org/ (accessed on 1 September 2024)) databases.

3. Results and Discussion

3.1. Effect of Different Temperature of EF on Sensory Quality of HJYBT

EF temperature significantly influenced the sensory quality of HJYBT (Figure 1). Optimal sensory characteristics in terms of liquor, appearance, and infused leaves (Figure 1A) were achieved at EF temperatures of 70 °C and 80 °C, resulting in a black and glossy appearance, red and clear liquor, and red infused leaves. Compared to unflavored tea (CK), EF improved the color, aroma, and appearance of tea leaves. However, EF temperatures above 90 °C led to a noticeable darkening of both the tea infusion and the leaves, with the most intense darkening occurring at T100. This change is likely due to a reduction in theaflavin content, an increase in theabrownins, and the caramelization of soluble sugars at higher temperatures [3,16,38]. Regarding taste (Figure 1B), CK exhibited a sweet and mellow flavor with a subtle green bitterness. Under low-temperature EF conditions, HJYBT presented a sweet, mellow, and refreshing taste. However, when the EF temperature exceeded 80 °C, the flavor shifted to sweet, mellow, and slightly astringent, with the quality of HJYBT particularly poor at T100. This deterioration likely results from the higher retention of amino acids and soluble sugars in black tea under lower temperature EF, which enhances sweetness and umami [39,40]. Notably, T80 exhibited the highest levels of sweetness and umami, and the lowest levels of bitterness and sourness, significantly differing from CK (p < 0.05). In terms of aroma profile (Figure 1C), the CK sample presented a mixture of various odors with the highest green and grassy notes. In contrast, T80 achieved an optimal aroma, characterized by a sweet and long-lasting floral scent. The increase in temperature likely accelerated Maillard and caramelization reactions in the tea leaves, enriching compounds such as pyrazine, pyrrole, and furfural [3,41]. When EF temperatures exceeded 90 °C, a pronounced burnt aroma developed, with HJYBT at T100 exhibiting caramel and roasted notes. Overall, HJYBT processed at an EF temperature of 80 °C demonstrated optimal quality, aligning with several previous studies [3,16,38].

3.2. Effect of Different EF on the Taste Components of HJYBT

3.2.1. Content Analysis of General Biochemistry Components

The taste quality of black tea is closely linked to its non-volatile metabolite (NVM) content. Spectrophotometry measured conventional biochemical components of HJYBT subjected to different EFs. Water extracts (WEs), representing the total soluble components, are a primary indicator of tea soup mellowness [42]. T80 exhibited the highest WE content at 42.43%, followed by T70 at 40.54%. Different EF temperatures significantly increased WE contents in HJYBT (Figure 2A). Tea polyphenols (TPs), which contribute to the bitter taste, were quantified as follows: CK, 20.81%; T60, 19.22%; T70, 16.43%; T80, 14.78%; T90, 14.41%; and T100, 14.16% (Figure 2B). Compared to CK, EF significantly reduced TPs in HJYBT, likely due to isomerization and polymerization under thermal physical chemistry, resulting in decreased content [3,43]. FAA, key umami NVMs, and SS, essential sweet NVMs in black tea [3], initially increased and then decreased. The FAA and SS content in T80 was notably higher than in CK and T100, while T100 exhibited lower levels than CK (Figure 2C,D). This pattern indicates that EF can enhance FAA and SS content, although excessively high EF temperatures reduce it, consistent with previous research [44].

TFs, TRs, and TBs play a significant role in determining the color and taste of black tea infusion. High levels of TFs and TRs, combined with a high TR/TF ratio and low TB content, are generally associated with superior tea quality [26]. TFs and TRs increased initially, peaking at T80 before declining. Various EFs significantly elevated TFs and TRs compared to CK (p < 0.05). The highest TR/TF ratio was also observed at T80 (Figure 2E,F). TB content rose significantly from CK to T100, reaching 4.03% at T100. The phenol–ammonia ratio, which indicates tea soup freshness, decreases as freshness improves, with T80 displaying a notably lower ratio of 3.25 compared to other samples. Overall, EF treatment reduced the bitterness of HJYBT soup, producing a mellower taste. However, EF temperatures above 80 °C substantially increased TB content, darkening the tea soup color and diminishing HJYBT quality, aligning with the sensory evaluation results (Figure 1A).

3.2.2. Content Analysis of Catechins, Flavones and Caffeine

Catechins, the primary NVMs influencing tea quality, contribute significantly to its astringent and bitter taste [45]. This study utilized HPLC to analyze eight catechin monomers (C, EC, GC, EGC, CG, ECG, GCG, and EGCG) to assess the impact of various EF temperatures on HJYBT quality (

Table 1. Contents of catechins and flavones measured by HPLC methods in the HJYBT samples (mean ± SD, mg/g).

Compounds	CK	T60	T70	T80	T90	T100
GA	3.09 ± 0.02 d	3.39 ± 0.10 c	3.61 ± 0.02 b	3.70 ± 0.03 b	3.90 ± 0.07 b	4.53 ± 0.01 a
GC	6.47 ± 0.07 b	7.15 ± 0.42 a	5.94 ± 0.60 b	5.23 ± 0.04 c	6.14 ± 0.61 b	3.95 ± 0.12 d
EGC	6.88 ± 0.24 a	6.41 ± 0.08 b	5.69 ± 0.09 c	5.13 ± 0.09 d	5.08 ± 0.05 d	4.03 ± 0.19 e
C	6.32 ± 0.03 c	6.68 ± 0.20 b	6.77 ± 0.04 b	7.17 ± 0.02 a	6.47 ± 0.25 bc	6.17 ± 0.05 d
CAF	39.17 ± 0.22 ab	38.89 ± 0.22 ab	40.37 ± 0.22 a	40.08 ± 0.76 a	39.05 ± 0.77 ab	40.88 ± 0.07 a
EGCG	11.71 ± 0.09 a	10.92 ± 0.06 b	10.62 ± 0.02 b	10.18 ± 0.02 bc	9.40 ± 0.10 c	8.16 ± 0.03 d
EC	3.14 ± 0.11 ab	2.99 ± 0.04 b	2.92 ± 0.01 b	3.23 ± 0.03 a	2.96 ± 0.02 b	2.87 ± 0.02 c
GCG	1.07 ± 0.06 b	1.01 ± 0.06 b	1.04 ± 0.04 b	1.03 ± 0.08 b	0.99 ± 0.05 bc	1.67 ± 0.02 a
ECG	2.96 ± 0.05 c	3.56 ± 0.22 a	3.70 ± 0.03 a	3.25 ± 0.16 b	3.56 ± 0.11 a	2.14 ± 0.01 d
CG	1.65 ± 0.20 cd	1.71 ± 0.07 c	1.81 ± 0.04 b	1.58 ± 0.03 d	1.88 ± 0.04 b	1.98 ± 0.08 a
Myr	2.53 ± 0.02 a	2.47 ± 0.02 a	2.14 ± 0.02 b	2.05 ± 0.03 c	2.27 ± 0.02 b	2.25 ± 0.02 b
Que	0.06 ± 0.00 c	0.07 ± 0.01 b	0.08 ± 0.01 a	0.07 ± 0.01 b	0.07 ± 0.01 b	0.08 ± 0.01 a
Lut	0.03 ± 0.00 b	0.04 ± 0.01 a	0.04 ± 0.00 a	0.04 ± 0.01 a	0.04 ± 0.01 a	0.04 ± 0.00 a
Kae	0.41 ± 0.02 a	0.42 ± 0.01 a	0.47 ± 0.01 a	0.44 ± 0.03 a	0.42 ± 0.04 a	0.39 ± 0.02 a
Rut	0.02 ± 0.00 c	0.02 ± 0.00 c	0.02 ± 0.00 c	0.03 ± 0.00 b	0.03 ± 0.00 b	0.04 ± 0.00 a
NECs	22.81 ± 0.42 ab	23.23 ± 0.41 a	21.32 ± 0.61 b	20.76 ± 0.05 b	20.65 ± 0.42 b	17.01 ± 0.24 c
ECs	17.40 ± 0.33 a	17.20 ± 0.11 a	17.17 ± 0.03 a	16.03 ± 0.15 b	15.83 ± 0.10 c	13.95 ± 0.11 d
TCs	40.21 ± 0.10 a	40.43 ± 0.51 a	38.49 ± 0.63 b	36.78 ± 0.20 c	36.48 ± 0.46 c	30.96 ± 0.18 d
TFS	3.05 ± 0.05 a	3.03 ± 0.04 a	2.75 ± 0.02 b	2.62 ± 0.03 bc	2.82 ± 0.07 b	2.81 ± 0.04 b

Notes: Different superscripts show significant differences in the same row (p < 0.05). GA, gallic acid; GC, gallocatechin; EGC, epigallocatechin; C, catechin; CAF, caffeine; EGCG, epigallocatechin gallate; EC, epicatechin; GCG, gallocatechin gallate; ECG, epicatechin gallate; CG, catechin gallate; Myr, myricetin; Que, quercetin; Lut, luteolin; Kae, kaempferol; Rut, rutin; NECs, non-ester catechins; ECs, ester catechins; TCs, total catechins; TFS: total flavones.

). The results indicated that EGCG was the most abundant catechin, followed by EGC. Generally, galloylated catechins are more bitter and astringent compared to their non-galloylated counterparts [46]. Galloylated catechins, including EGCG, ECG, and GCG, are key polyphenols in green tea responsible for its bitterness and astringency [43]. Among the four detected galloylated catechins, EGCG exhibited higher abundances, with their concentrations fluctuating and reaching minimum and maximum values at T100 and CK, respectively. A previous study reported that EGCG, a major catechin derivative in HJYBT extract, reduced bile acid reabsorption, decreased intestinal bile acids levels, and then inhibited lipid absorption in mice [24,47]. Its levels were still high after EF at 60 °C, 70 °C, and 80 °C. Non-epicatechins such as CG and GCG significantly increased during EF treatment, likely due to the high temperatures causing the isomerization of EGC, EGCG, and ECG into GCG and CG [1]. Additionally, the bitterness and astringency of epicatechins are more pronounced than those of non-epicatechins at equivalent levels [1,46]. Non-galloylated catechins like EGC and EC are closely associated with the sweet aftertaste of tea infusion [1]. Notably, the differences in C and EC contents before and after EF treatment were insignificant. In contrast, the total catechins (TCs) concentrations significantly decreased (p < 0.05) post-EF treatment, reaching a minimum of 30.96 mg/g in T100.

GA, identified as an umami-enhancing compound in green tea, increased during the EF procedure, peaking at the T100 stage at approximately 1.5 times the CK level [45]. Bitterness and astringency in tea infusion were also affected by compounds like caffeine, flavonol, and flavonol glycosides [48,49]. The primary alkaloids in tea, including caffeine (CAF), theophylline, and theobromine, contribute to these sensory characteristics, with CAF being particularly abundant and significantly enhancing astringency and bitterness at varying concentrations [46]. After EF treatment, the content of CAF remained almost unchanged. Among the four detected flavonols, Myr showed relatively higher abundances, with concentrations fluctuating, reaching their minimum at T80 and maximum at CK. Similarly, total flavonols (TFS) exhibited fluctuations during EF processing, peaking at CK and reaching a minimum at T80. These results suggest that EF can reduce the bitterness and astringency of HJYBT infusion to some extent.

3.2.3. Analysis of the Content of FAAs

FAAs are essential chemical constituents of tea, with their concentration, composition, and degradation status significantly influencing tea quality, particularly enhancing its freshness and aroma [43,45]. HPLC quantified the contents of 20 FAAs in each sample (

Table 2. Contents of free amino acids measured by HPLC methods in the HJYBT samples (mean ± SD, mg/g).

Compounds	CK	T60	T70	T80	T90	T100
Asp	3.28 ± 0.01 a	3.00 ± 0.05 b	2.54 ± 0.00 c	2.75 ± 0.08 c	2.18 ± 0.02 d	2.02 ± 0.01 d
Glu	0.40 ± 0.00 a	0.42 ± 0.00 a	0.44 ± 0.00 a	0.33 ± 0.03 b	0.32 ± 0.06 b	0.32 ± 0.06 b
Gln	1.35 ± 0.00 a	1.16 ± 0.00 b	0.62 ± 0.00 e	0.73 ± 0.05 d	0.92 ± 0.02 c	0.65 ± 0.03 e
Thea	9.01 ± 0.03 a	8.2 ± 0.10 b	7.82 ± 0.04 c	6.95 ± 0.07 d	6.08 ± 0.08 e	5.16 ± 0.03 f
His	0.20 ± 0.00 a	0.17 ± 0.00 a	0.18 ± 0.01 a	0.18 ± 0.01 a	0.16 ± 0.01 ab	0.16 ± 0.01 ab
Arg	1.56 ± 0.03 a	1.58 ± 0.01 a	1.57 ± 0.02 a	1.41 ± 0.09 b	1.26 ± 0.02 c	1.13 ± 0.00 c
Tyr	0.50 ± 0.01 a	0.50 ± 0.00 a	0.52 ± 0.00 a	0.44 ± 0.03 b	0.44 ± 0.02 b	0.44 ± 0.02 b
Val	1.36 ± 0.00 a	1.21 ± 0.01 b	1.25 ± 0.02 b	1.14 ± 0.05 c	1.04 ± 0.08 c	0.84 ± 0.03 d
Phe	0.98 ± 0.00 a	0.93 ± 0.00 a	0.88 ± 0.00 a	0.73 ± 0.01 b	0.66 ± 0.02 b	0.47 ± 0.01 c
Lys	0.81 ± 0.01 a	0.83 ± 0.01 a	0.86 ± 0.01 a	0.7 ± 0.06 b	0.7 ± 0.03 b	0.46 ± 0.01 c
Leu	1.21 ± 0.01 a	1.11 ± 0.01 a	1.13 ± 0.01 a	0.92 ± 0.05 b	0.85 ± 0.01 b	0.59 ± 0.00 c
Ser	1.08 ± 0.00 a	0.9 ± 0.01 a	0.96 ± 0.02 a	0.78 ± 0.05 b	0.79 ± 0.02 b	0.56 ± 0.03 c
Gly	0.14 ± 0.00 a	0.14 ± 0.00 a	0.15 ± 0.01 a	0.14 ± 0.00 a	0.12 ± 0.00 b	0.12 ± 0.00 b
Thr	0.81 ± 0.00 a	0.66 ± 0.00 b	0.69 ± 0.00 b	0.57 ± 0.04 c	0.57 ± 0.03 c	0.52 ± 0.05 c
Ala	0.44 ± 0.00 a	0.45 ± 0.00 a	0.49 ± 0.00 a	0.40 ± 0.03 a	0.41 ± 0.03 a	0.41 ± 0.03 a
Cys	0.07 ± 0.00 b	0.05 ± 0.00 d	0.06 ± 0.00 c	0.11 ± 0.00 a	0.07 ± 0.01 b	0.07 ± 0.01 b
L-asp	4.31 ± 0.06 a	4.42 ± 0.06 a	4.5 ± 0.06 a	3.92 ± 0.10 b	3.68 ± 0.09 b	3.16 ± 0.01 c
Met	0.19 ± 0.01 a	0.19 ± 0.00 a	0.21 ± 0.01 a	0.17 ± 0.02 b	0.19 ± 0.06 a	0.19 ± 0.06 a
Try	0.83 ± 0.02 a	0.82 ± 0.01 a	0.84 ± 0.01 a	0.75 ± 0.04 b	0.7 ± 0.02 b	0.54 ± 0.02 c
Iso	0.88 ± 0.02 a	0.89 ± 0.01 a	0.92 ± 0.02 a	0.75 ± 0.06 b	0.73 ± 0.02 b	0.55 ± 0.02 c
TFAAs	29.42 ± 0.18 a	27.64 ± 0.27 b	26.61 ± 0.16 b	23.86 ± 0.73 c	21.87 ± 0.38 c	18.35 ± 0.1 d

Notes: Different superscripts show significant differences in the same row (p < 0.05). Asp, asparagic acid; Glu, glutamic acid; Gln, glutamine; Thea, theanine; His, histidine; Arg, arginine; Tyr, tyrosine; Val, valine; Phe, phenylalanine; Lys, lysine; Leu, leucine; Ser, serine; Gly, Glycine; Thr, threonine; Ala, alanine; Cys, cystine; L-asp, L-asparagine; Met, methionine; Try, tryptophan; Iso, isoleucine; TFAAs, total free amino acids.

). These amino acids can be classified by taste: umami (Asp, Glu, Gln, and Thea), bitter (Leu, Val, Arg, Phe, Lys, His, and Thr), or sweet (Thr, Gly, Ser, Ala, Cys, and Pro) [50]. Thea is the most abundant, followed by L-asp, Asp, and Arg. The results indicated that all amino acids, especially Thea, Gln, Val, Ser, Thr, and Val, decreased, with some imparting umami or sweet flavors to the tea infusion. For instance, Thea and Gln are known as umami-enhancing compounds in matcha, a high-grade powdered green tea from Japan [45,51]. Ser, Ala, Gly, and Thr enhance the sweetness of black tea infusion [52]. Additionally, the reduced amino acids might act as aroma precursors, contributing to the aromatic quality of HJYBT infusion through the Maillard reaction [53]. For instance, Thea produces a strong caramel aroma, Glu generates floral notes, and Thr and Ser yield wine-like aromas. Furthermore, Phe in tea can be transformed into distinct flowery or rose-like volatiles [53,54]. These findings suggest that amino acid degradation contributes to the formation of sweet or floral aromas in HJYBT, aligning with sensory evaluation results.

3.2.4. Multivariate Statistical Analysis and Screening for Differential NVMs

A non-supervised metrology PCA tool analyzed the detected NVMs across different HJYBT samples (Figure 3A). The first principal component (PC1) accounted for 59.5% of the total variance, while the second principal component (PC2) accounted for 16.0%. The clustering of replicate samples treated with various EFs indicated significant differentiation, demonstrating data reliability. CK, T60, T70, T80, T90, and T100 were distinctly separated from each other. Specifically, CK and T60 were located in the upper right quadrant, T70 in the lower right quadrant, T80 and T90 in the lower left quadrant, and T100 in the upper left quadrant, indicating significant differences before and after EF treatment. In the PC1 direction from right to left (Figure 3A), EF values were regulated by temperatures from 60 °C, 70 °C, 80 °C, and 90 °C to 100 °C, suggesting that EF plays a key role in the evolution of taste components. To better visualize and understand the chemical changes during EF of HJYBT, hierarchical clustering analysis (HCA) was performed. As shown in Figure 3B, HCA distinctly separated the samples by EF treatment. Additionally, the six HJYBT samples were divided into clusters I and II, with CK, T60, and T70 assigned to cluster I, T80, T90, and T100 assigned to cluster II. The HCA tree structure aligned with PCA and sensory analysis data.

OPLS-DA, a supervised multivariate statistical analysis method, filters orthogonal variables unrelated to categorical variables, thereby providing deeper insights into the impact of EF on NVMs [33]. To further elucidate the changes in NVMs during EF treatment, 48 identified NVMs served as independent variables, while 6 treatment groups acted as dependent variables in the OPLS-DA (Figure 3C). This analysis effectively distinguished all HJYBT samples from one another. Cross-validation with 200 permutation tests revealed negative intercepts of Q2, indicating the reliability of the OPLS-DA models (Figure S1).

Based on the OPLS-DA model, differential NVMs were identified using variable importance in projection (VIP) and fold change values across the stages (CK vs. T60, T60 vs. T70, T70 vs. T80, T80 vs. T90, and T90 vs. T100). Fifteen differential NVMs were selected with VIP > 1, including WEs, TFAA, TPs, TCs, NECs, CAF, GC, ECs, EGCG, Thea, ECG, TRs, Par, EGC, and EC (Figure S1). Correlations between taste data from descriptive sensory analysis and differential NVMs were analyzed using partial least squares discriminant analysis (PLS-DA, Figure 3D). The quality parameters of the PLS-DA model indicated a robust fit (R2Y = 0.985) and strong predictive ability (Q2 = 0.968). Figure 3D illustrates the correlation between 15 NVMs and sensory characteristics. WEs and ECG showed a strong association with mellow attributes, while sourness and astringency were closely linked to Par and TRs. Additionally, GC was related to sweetness and umami. Most of the other NVMs were associated with bitterness, except for EC, which does not fully align with previous studies [45,46,48,55]. This discrepancy may be due to the dynamic and complex changes in substances during the EF processing of HJYBT. Analyzing only the main ingredients is insufficient to fully understand their impact on quality. Consequently, ongoing targeted metabolomics studies using LC-MS are investigating the effects of different EF temperatures on HJYBT metabolites.

3.3. Effect of Different EFs on the Aroma and Volatile Compounds of HJYBT

3.3.1. Results of E-Nose Analysis

The E-nose demonstrates a high sensitivity to odor information, with slight variations in volatile compounds producing distinct sensor responses [56]. Currently, E-nose technology is used to evaluate black tea quality grades [29], monitor aroma changes during the roasting of oolong tea [57], and distinguish aroma differences in tricholoma matsutake from different origins [56]. Figure 4 illustrates the response of 10 E-nose sensors to HJYBT at various EF temperatures. Higher response values indicated greater concentrations of the detected components. The E-nose responses for the six HJYBT samples exhibited similar profiles, though the intensity of responses from the 10 metal oxide sensors varied among the samples. The strongest responses were observed in the W1W, W2W, W5S, and W5C receptors, suggesting varying levels of terpenes, aromatic compounds, sulfur and chlorine compounds, nitrogen oxides, and hydrocarbons among the samples. Conversely, the lowest response values were detected in W1C, W3C, and W6S across all samples. Nitrogen oxides, associated with Maillard-related products, showed a significantly higher response in the W5S sensor for HJYBT processed at 80 °C, likely due to the accumulation of nitrogen oxides from the Maillard reaction at increased EF temperatures [33]. These results indicated that the aroma profile of HJYBT varied significantly with EF processing.

3.3.2. Overview of the Profile of Volatile Compounds (VMs)

To clarify the impact of EF on the aroma quality of HJYBT, non-targeted metabolomics using GC-TOFMS measured and analyzed aroma compounds across various EF temperatures. Correlation analysis among samples confirmed the high reliability and repeatability of the obtained data (Figure S2). A total of 488 VMs were identified by SAFE/GC-MS, and their relative contents were calculated (Table S1). Based on their functional groups, the VMs were categorized as acids, alcohols, aldehydes, amines, aromatics, esters, ethers, heterocyclic compounds, hydrocarbons, ketones, phenols, terpenoids, and others (Figure 5A and Table S1).

Hydrocarbons (96 compounds) constituted 19.6% of the total VMs, followed by esters (93), alcohols (77), ketones (47), aldehydes (45), aromatics (32), terpenoids (24), acids (12), amines (8), and phenols (6), comprising 19.1%, 16.0%, 9.6%, 9.0%, 6.6%, 4.9%, 2.5%, 1.6%, and 1.2%, respectively. These ten volatile types accounted for approximately 90% of the total VMs. Additionally, five ethers were identified, representing 1.0% of the total VMs, consistent with other black tea processing studies [2,21,33,58]. Detailed information on VMs in HJYBT across various EF stages is presented in Figure 5A and Table S1.

Figure 5B,C depict the variations in the total VMs of HJYBT during EF. The total VMs exhibited a dynamic pattern under different EF temperatures, initially decreasing, then increasing, and subsequently decreasing again. The peak value of 73,600.23 µg/kg occurred at T80, indicating that both excessively low and high EF temperatures affect the aroma of HJYBT. Notably, while the relative content of aroma compounds in CK was higher than in other samples except T80, its sensory aroma quality was comparatively weaker, likely due to imbalances in component ratios. The proportion of specific volatiles within the total volatile content showed significant variation. For instance, alcohols constituted 32.61% of CK but decreased to 19.36% at T100. Conversely, heterocyclic compounds, which accounted for 8.25% in CK, increased with EF temperature, peaking at 10.62% in T100. Alcohols were among the most prevalent VMs in HJYBT, comprising 19.36–38.33% of the total VMs, followed by aldehydes, hydrocarbons, esters, and heterocyclic compounds, collectively representing over 85% of the total VMs. In contrast, terpenoids, acids, aromatics, ketones, and phenols were present at comparatively lower levels. Figure 5D provides the contents of the top 20 VMs in HJYBT, which accounted for 55.69% to 75.68% of the total volatiles. Among these, fifteen VMs, including geraniol, 2-phenylethanol, phenylacetaldehyde, benzyl alcohol, linalool oxide, linalool, hexanal, 3-methylbutyraldehyde, heptanal, methyl salicylate, (E)-2-hexenal, benzaldehyde, 2-pentyl-furan, nonanal, and (Z)-2-penten-1-ol, have been identified as key odorants in the four most famous black teas [2,21,59,60]. Consequently, these VMs in HJYBT, which constitute a substantial proportion of the total VMs, significantly contribute to the tea’s aroma and warrant further attention.

Alcohols were the most abundant VMs (20,148.84 µg/kg) in HJYBT before EF, constituting over 30% of the total VMs content. The concentration of alcohols decreased to 17,628.90 μg/kg in T60, then increased during subsequent EF treatment. The peak value occurred at T80, followed by a sharp decline, reaching 10,398.18 μg/kg by T100. In tea, alcohols are unstable and readily oxidize to aldehydes and/or acids, as observed in Wuyi rock tea during full fire processing [11]. Aldehydes and terpenoids displayed a similar trend to alcohols, peaking at T80 and significantly exceeding CK (p < 0.05). In contrast, the content and proportion of amines and phenols were lowest at T80, significantly lower than CK (p < 0.05). A significant number of heterocyclic compounds, such as pyrazine, pyrrole, and pyran structures, commonly form through the Maillard reaction. These compounds were found in oolong tea subjected to full firing processing but not in unroasted oolong tea [11,61]. In this study, the content of heterocyclics in HJYBT consistently increased during EF processing, peaking at 5700.98 μg/kg in T100. Additionally, a transformation in odor type was observed, with a predominant floral flavor in T70 and T80 samples, shifting to caramel and roasted aromas in T100. Similarly, the contents of ketones and hydrocarbons increased significantly, by 1.60 and 2.71 times from CK to T100, respectively.

3.3.3. Variance Analysis of VMs in HJYBT during EF Processing

Multivariate statistical analysis based on GC-TOF/MS was conducted on all samples from different EF stages to better reflect changes in HJYBT quality during processing. The PCA score plot (Figure S3A) shows that the first principal component (PC1) and the second principal component (PC2) accounted for 51.2% and 22.2%, respectively, with a total contribution rate of 73.4%. This effectively highlights significant differences in volatile compounds among HJYBT samples subjected to various EF processes. The HCA results (Figure S3B) aligned with those of the PCA. Overall, the unsupervised methods’ results were consistent with sensory evaluation, providing a solid foundation for further analysis.

To further analyze the general tendency and identify contributing differential VMs, the 488 volatiles of different EF of HJYBT were examined using OPLS-DA. Figure 6A shows the clear separation of samples before and after EF processing along the x-axis of the score plot, indicating significant differences in metabolic composition. This suggests that the EF procedure greatly influences the sensory quality of HJYBT. To further investigate the impact of the EF process on the aroma substances of HJYBT, all tea samples were divided into two groups (before EF and after EF) and subjected to OPLS-DA analysis, as shown in Figure 6B. The pre-EF samples are positioned on the left side of the diagram, while the post-EF samples are on the right side, demonstrating that the EF process significantly affects the aroma quality of HJYBT. To elucidate the impact of different EF temperatures on the aroma substances of HJYBT, an OPLS-DA model was constructed based on samples redivided into four groups (T60, T70, T80, T90, and T100) (Figure 6C). The OPLS-DA score plots revealed that T60 and T70, as well as T90 and T100, clustered together, while T80 was distinctly separated along the x-axis. This indicates a continuous transformation of the metabolic components in HJYBT with increasing EF temperature, with 80 °C potentially being a critical turning point for HJYBT quality formation. Notably, the R2Y and Q2 scores in these analyses exceeded 0.95, demonstrating high explanatory power and predictive accuracy of the model. The reliability of each OPLS-DA model was confirmed through permutation tests with 200 random iterations, where Y-intercepts of Q2 were all less than 0, indicating no overfitting [2,62].

To elucidate the effect of EF on HJYBT aroma attributes, differential VMs in the six key modules were analyzed via OPLS-DA of CK vs. T60, CK vs. T70, CK vs. T80, CK vs. T90, and CK vs. T100, based on the criteria of variable importance values (VIP) > 1 and p < 0.05 [3] (Figure S4). A total of 88 differential NVMs were identified, with 54 VMs observed between CK and T60 (22 upregulated and 32 downregulated), 54 VMs between CK and T70 (26 upregulated and 28 downregulated), 47 VMs between CK and T80 (26 upregulated and 21 downregulated), 59 VMs between CK and T90 (35 upregulated and 24 downregulated), and 66 VMs between CK and T100 (34 upregulated and 32 downregulated) (Figure 6D). The EF process significantly influences the aroma quality of HJYBT. When the EF temperature exceeded 80 °C, higher temperatures led to more pronounced changes in VMs, resulting in a greater number of differential VMs. Increased temperature accelerated Maillard and caramelization reactions in the leaves, enriching heterocyclic compounds such as pyrazine, pyrrole, and furfural [3,41].

A Venn diagram (Figure 6E) identified common and unique differential VMs among various tea samples. Comparisons between CK and T60, T70, T80, T90, and T100 revealed 33 common VMs, including 9 alcohols, 7 hydrocarbons, 5 aldehydes, 4 esters, 3 heterocyclic compounds, 2 terpenoids, 1 ketone, 1 ether, and 1 acid. To better visualize the changes in differential VMs, a heatmap analysis combined with hierarchical clustering was conducted on the 33 identified differential VMs (Figure 6F and

Table 3. A total of 33 differential VMs of HJYBT at different EF processing temperatures.

No.	Compounds	CK vs. T60			CK vs. T70			CK vs. T80			CK vs. T90			CK vs. T100
		VIP	p-Value	Type	VIP	p-Value	Type	VIP	p-Value	Type	VIP	p-Value	Type	VIP	p-Value	Type
Alcohols
1	Geraniol	1.25	0.02	Up	2.46	0.01	Up	4.48	0.00	Up	1.27	0.00	Up	3.16	0.01	Down
2	2-phenylethanol	2.83	0.04	Up	5.71	0.00	Up	7.47	0.00	Up	3.49	0.00	Up	2.62	0.00	Down
3	1-penten-3-ol	2.46	0.04	Down	2.56	0.01	Down	2.46	0.00	Down	3.25	0.00	Down	2.92	0.00	Down
4	(Z)-2-penten-1-ol	4.34	0.02	Down	4.34	0.01	Down	3.51	0.00	Down	4.41	0.00	Down	3.97	0.00	Down
5	(E)-3-hexen-1-ol	4.98	0.00	Down	4.53	0.00	Down	3.81	0.00	Down	4.51	0.00	Down	4.25	0.00	Down
6	1-hexanol	3.60	0.01	Down	3.08	0.00	Down	2.65	0.00	Down	3.21	0.00	Down	3.25	0.00	Down
7	(E)-2-hexen-1-ol	2.87	0.00	Down	2.58	0.00	Down	2.15	0.00	Down	2.34	0.00	Down	2.45	0.00	Down
8	1-octanol	1.69	0.00	Down	1.34	0.00	Down	1.18	0.00	Down	1.52	0.00	Down	1.49	0.00	Down
9	α-terpineol	1.47	0.00	Up	1.23	0.01	Up	2.87	0.00	Up	2.85	0.00	Up	4.02	0.00	Up
Hydrocarbons
10	Dodecane	1.64	0.03	Down	5.90	0.00	Up	5.28	0.00	Up	6.70	0.00	Up	7.27	0.00	Up
11	Tridecane	2.40	0.04	Up	3.03	0.01	Up	2.43	0.00	Up	2.97	0.01	Up	2.98	0.00	Up
12	Decane	2.88	0.00	Down	2.35	0.00	Down	1.06	0.02	Down	2.22	0.00	Down	1.20	0.00	Down
13	3-methyl-tridecane	1.91	0.00	Up	1.48	0.00	Up	1.39	0.00	Up	1.60	0.00	Up	1.84	0.00	Up
14	3-methylene-undecane	1.12	0.01	Down	1.21	0.00	Up	1.030	0.00	Up	1.16	0.00	Up	1.41	0.00	Up
15	Linalool	2.32	0.00	Up	2.57	0.00	Up	2.78	0.00	Up	2.16	0.00	Down	1.87	0.00	Down
16	Tetradecane	1.39	0.00	Up	1.27	0.00	Up	1.00	0.00	Up	1.56	0.00	Up	1.43	0.00	Up
Aldehydes
17	Heptanal	2.79	0.01	Down	3.05	0.00	Down	3.14	0.00	Down	4.90	0.00	Down	4.88	0.00	Down
18	Pentanal	3.18	0.00	Up	2.94	0.00	Up	2.66	0.00	Up	2.41	0.00	Up	2.19	0.00	Up
19	Nonanal	2.51	0.00	Up	2.09	0.00	Up	1.94	0.00	Up	1.67	0.00	Down	1.82	0.00	Down
20	Furfural	1.49	0.00	Up	1.26	0.00	Up	187	0.00	Up	4.05	0.00	Up	4.74	0.00	Up
21	α-citral	1.50	0.00	Down	1.29	0.00	Down	1.10	0.00	Down	1.31	0.00	Down	1.26	0.00	Down
Esters
22	Methyl salicylate	7.66	0.00	Down	6.52	0.00	Down	5.51	0.00	Down	6.31	0.00	Down	6.00	0.00	Down
23	γ-Decalactone	5.46	0.00	Down	4.45	0.00	Down	3.71	0.00	Down	4.47	0.00	Down	3.98	0.00	Down
24	Allyl phenoxyacetate	2.40	0.00	Down	1.86	0.00	Down	1.56	0.00	Down	1.88	0.00	Down	1.69	0.00	Down
25	Hexanoic acid, hexyl ester	2.53	0.00	Down	1.83	0.00	Down	1.49	0.00	Down	1.78	0.00	Down	1.57	0.00	Down
Heterocyclic compounds
26	2-pentyl-furan	2.53	0.00	Up	1.97	0.00	Up	1.71	0.00	Up	2.50	0.00	Up	1.92	0.00	Up
27	Pyrrole	2.11	0.00	Up	1.24	0.00	Up	2.01	0.00	Up	2.31	0.00	Up	1.18	0.00	Up
28	Pyridine	1.21	0.00	Up	1.32	0.00	Up	1.81	0.00	Up	2.43	0.00	Up	3.12	0.00	Up
Terpenoids
29	Myrcene	1.56	0.01	Up	1.39	0.00	Up	1.35	0.00	Up	2.12	0.00	Up	3.04	0.00	Up
30	2,4-dimethyl-1-heptene	1.03	0.00	Up	1.04	0.00	Up	1.40	0.00	Up	1.01	0.00	Up	1.29	0.00	Up
Ketones
31	(E)-β-ionone	2.13	0.00	Up	2.21	0.00	Up	1.98	0.00	Up	3.10	0.00	Up	2.64	0.00	Up
Ethers
32	Diphenyl ether	5.99	0.00	Down	4.49	0.00	Down	3.75	0.00	Down	4.51	0.00	Down	4.02	0.00	Down
Acid
33	Acetic acid	3.22	0.00	Down	2.46	0.00	Down	2.29	0.00	Down	2.85	0.00	Down	2.53	0.00	Down

Note: VIP, variable important in projection values were extracted from the OPLS-DA result; p-value, p-values were ANOVA results; Up, upregulated, Down, downregulated.

). These VMs were categorized into four groups, each displaying distinct trends. In group 1, VM concentrations increased after T60, peaking after EF processing at 100 °C. In group 2, volatile VMs increased rapidly after T80 and then swiftly declined. Apart from decane, VMs concentrations in groups 3 and 4 decreased sharply after EF processing and remained low throughout the subsequent EF stages. These selected VMs from HJYBT exhibited distinct trends during EF processing, indicating their potential as markers for differentiating EF temperature variations in HJYBT. Additionally, the main volatiles responsible for the sweet, floral, and fruity aromas of black tea, such as geraniol, 2-phenylethanol, (E)-β-ionone, nonanal, γ-decalactone, and α-citral, were identified (Table 3). Further research aimed to pinpoint key VMs contributing to the aroma quality differences in HJYBT at various EF stages.

3.3.4. Key Aroma Compounds in HJYBT during EF Processing

The aroma quality of tea leaves is intricately connected to the composition and proportion of volatile aroma substances, with their overall influence determined by concentration and odor threshold [11]. The odor activity value (OAV) is used to evaluate the contribution of volatile compounds to tea aroma, where VMs with an OAV greater than one are considered to positively impact the overall aroma profile [2,11]. During HJYBT EF processing, 19 VMs exhibited OAVs greater than one (

Table 4. The selected volatiles with OAVs above one in HJYBT EF processing.

No.	Compounds	CAS	Odor Type	Threshold (μg/kg−1)	OAVs
					CK	T60	T70	T80	T90	T100
1	Geraniol	106-24-1	Floral	7.5 [11]	566.17	647.07	772.32	1183.36	655.23	362.57
2	Heptanal	111-71-7	Green	3 [32]	922.04	821.27	691.80	553.49	288.55	110.58
3	Methyl salicylate	118-61-6	Floral	40 [6]	123.15	69.03	44.94	37.88	44.03	31.09
4	2-pentyl-furan	3777-69-3	Fruity	4.8 [32]	116.79	156.29	217.59	226.02	260.58	236.24
5	Pentanal	110-62-3	Nutty	400 [32]	0.64	1.57	2.23	2.62	1.82	1.87
6	Nonanal	124-19-6	Floral	10 [11]	55.94	79.59	88.41	98.53	33.02	21.49
7	(Z)-2-penten-1-ol	1576-95-0	Green	720 [11]	3.18	2.14	1.06	1.23	1.02	0.93
8	(E)-3-hexen-1-ol	928-97-2	Green	110 [11]	21.74	13.24	7.90	6.85	6.67	4.88
9	1-hexanol	111-27-3	Green	5.6 [11]	265.97	177.66	138.55	123.09	117.75	71.28
10	Furfural	1998-1-1	Roasted	770 [11]	0.17	0.28	0.33	0.68	1.86	3.15
11	(E)-β-ionone	79-77-6	Floral	7 [32]	3.53	10.79	14.89	28.37	14.77	11.12
12	Myrcene	123-35-3	Woody	15 [11]	8.73	15.10	18.45	22.47	32.56	72.05
13	Linalool	78-70-6	Floral	10 [11]	116.68	190.22	214.61	350.62	80.03	43.49
14	(E)-2-hexen-1-ol	928-95-0	Green	5.6 [11]	136.63	81.34	48.56	43.29	58.40	26.46
15	1-octanol	111-87-5	Chemical	125.8 [56]	2.60	1.75	1.53	1.35	1.14	0.78
16	γ-decalactone	706-14-9	Fruity	110 [11]	15.50	5.56	2.29	1.46	1.13	0.79
17	α-citral	141-27-5	Green	32 [11]	6.35	3.76	2.55	2.12	2.10	1.29
18	Hexanoic acid, hexyl ester	6378-65-0	Green	70 [6]	3.90	0.53	0.37	0.31	0.27	0.25
19	2-ethyl-6-methyl-pyrazine	13925-03-6	Roasted	40 [32]	0.10	0.27	1.12	1.95	3.08	6.21

). Most of these compounds exhibited floral, fruity, green, roasted, and caramel odors, significantly contributing to the overall aroma profile of HJYBT. Among them, 12 compounds had an OAV greater than 10, indicating their substantial impact on HJYBT’s dominant aroma profile. These include geraniol, linalool, heptanal, methyl salicylate, 2-pentyl-furan, nonanal, (E)-3-hexen-1-ol, 1-hexanol, (E)-β-ionone, myrcene, and (E)-2-hexen-1-ol. Notably, heptanal, which imparts a green-like odor, had the highest OAV of 922.05 in CK. Its relative content and OAV decreased progressively with rising EF temperatures, reaching the lowest level at T100, only 12% of CK (Table 4 and Figure 7). This reduction likely contributes to the loss of HJYBT’s green odors after EF.

Several alcohols, including (Z)-2-penten-1-ol, (E)-3-hexen-1-ol, 1-hexanol, and (E)-2-hexen-1-ol, were identified as key odorants contributing to green odors and exhibited higher OAVs in CK compared to other samples. The OAVs in CK were 3.42–5.16 times higher than those in T100 (Table 4), making these volatiles significant contributors to the reduced green odor of HJYBT post-EF processing. Geraniol, associated with the floral odor of black tea [62], reached its maximum content and OAV at T80, which may account for the more pronounced floral aroma in this sample. Additionally, several fatty aldehydes with OAVs > 1, including pentanal, nonanal, furfural, and α-citral, were detected in HJYBT. Nonanal, with a floral odor and the highest OAV of 98.53 in T80, likely had a positive impact on HJYBT’s aroma. In contrast, the increased furfural and pentanal in T100 exhibited roasted and nutty flavors, negatively impacting the aroma quality of black tea [21,33]. Methyl salicylate and γ-decalactone, key odorants presenting floral and fruity notes, had OAVs of 123.15 and 15.50 in CK, respectively, which were 3.96 and 19.50 times higher than those in T100, indicating that higher EF temperatures reduce the aroma quality of HJYBT. Additionally, the OAVs of 2-pentyl-furan (fruity), (E)-β-ionone, and linalool (floral) significantly increased when EF temperature reached 80 °C. (E)-β-ionone and linalool substantially contributed to the floral notes of black tea [7,21,33]. Overall, floral and fruity volatiles showed higher contents and OAVs, while green notes decreased after EF processing, aligning with the sensory evaluation results (Figure 1). Specifically, temperatures exceeding 80 °C led to a decrease in floral and fruity aromas, replaced by caramel and high-fire flavors, diminishing the overall quality of HJYBT.

3.3.5. Correlation Analysis of Key Differential VMs and QDA Indicators Based on PLS-DA

To evaluate the relationship between aroma indicators (floral, fruity, sweet, green and grassy, honey, caramel, and roast) and the 19 key differential VMs, data from Figure 1B and Table 4 were correlated using PLS-DA. The quality parameters of the PLS-DA model indicated a strong fit (R2Y = 0.980) and robust predictive ability (Q2 = 0.903). Figure 8 illustrates the correlation between the 19 key differential VMs and sensory characteristics, with the size of the green circle representing OAV. The compounds 2-ethyl-6-methyl-pyrazine, furfural, and myrcene were linked to caramel and roast attributes. Similarly, geraniol, linalool, nonanal, and (E)-β-ionone were closely associated with floral, fruity, sweet, and honey attributes, consistent with previous studies [2,21]. Notably, geraniol and linalool, represented by large green circles, were nearer to floral and fruity attributes. Both compounds had higher relative concentrations in T80 compared to other samples, potentially contributing to the floral and fruity aroma at T80. The remaining VMs were more closely related to green and grass attributes. The PLS-DA results further corroborate the qualitative and quantitative analyses and the sensory analysis of aroma-active compounds in HJYBT.

4. Conclusions

In the present study, EF processing significantly influenced the sensory quality, objective quantification indicators, and VMs conversion of HJYBT. Compared to CK, EF treatment enhanced the sweetness and umami of HJYBT, improved floral, fruity, and sweet aromas, and reduced bitterness and green odors. An EF temperature of 80 °C yielded optimal umami, sweetness, and overall taste, with heightened aromatic intensity, particularly in floral and fruity notes among the five EF temperatures. Spectrophotometry and HPLC analysis identified 15 differential NVMs with VIP > 1 between pre- and post-EF processing, including WEs, TFAA, TPs, TCs, and NECs. EF processing increased levels of WEs and TRs, while decreasing levels of TPs, TCs, ECs, EGCG, ECG, and CAF, thereby enhancing sweetness, umami, and overall taste, and reducing astringency and bitterness. Nineteen VMs were identified as significant variables influencing the aroma variations of HJYBT during different EF processing stages, all with OAVs over 1. Geraniol and linalool contribute to floral and fruity odors, while 2-ethyl-6-methyl-pyrazine, furfural, and myrcene contribute to caramel and roast odors. Other key VMs, such as heptanal, methyl salicylate, α-citral, 1-hexanol, and (E)-3-hexen-1-ol, may affect the green and grassy odors of HJYBT. Excessive EF processing temperatures above 80 °C lead to the degradation of sweetness, umami taste, and floral and fruity odor metabolites, resulting in decreased HJYBT quality. Therefore, 80 °C is considered the optimal EF processing temperature for HJYBT (Figure 9). These results provide valuable insights for improving HJYBT quality during EF processing. Future research will focus on the effects of varying EF times and methods on HJYBT quality.