Airflow: A double-edged sword in shaping white tea quality

Abstract

Airflow significantly impacts white tea quality by regulating moisture loss and physiochemical reactions. LF-NMR, TEM, SEM, GC-TOF-MS, and UPLC-TQS-MS were used to compare the impacts of withering with different airflows, 0 m/s (AF0), 0.5 m/s (AF0.5), and 1 m/s (AF1), on white tea quality. AF0 impeded moisture loss, accumulating green aroma compounds like (E)-2-hexen-1-ol, 1-hexanol, and astringents including catechins and caffeine. AF0.5 and AF1 expedited stomatal opening, facilitating water migration from stems to leaves or buds and converting water states to bound water. Water stress increased cellular damage, enzymatic activity and physiochemical reactions. AF0.5 enhanced floral aroma, a mellow, fresh taste, with grayish-green color transformation by forming linalool, its oxide, nerol, organic acids, and amino acids. Besides floral characteristics, AF1 yielded grassy notes with 3-hexen-1-ol. Research indicates that AF0.5 optimizes white tea quality by balancing oxidation, hydrolysis, and synthesis, offering insights into white tea processing and withering equipment parameters.

Highlights

• •Airflow acts as a double-edged sword regulating white tea quality.
• •Airflow impacts physicochemical reactions via water loss and withering regulation.
• •Airflow promotes stomatal opening, cell damage, and increased enzyme activity.
• •Both high and low airflow rates can cause grassy aromas to accumulate in white tea.
• •Moderate airflow at 0.5 m/s results in floral, mellow, and refreshing white tea.

1. Introduction

White tea (Camellia sinensis), a traditional tea category originating from China and characterized by its rich floral aroma, sweet taste, and numerous health benefits, is produced primarily in Fujian Province. The unique processing method for white tea involves withering and drying, and the withering stage is pivotal in shaping its quality characteristics (Yue et al., 2019). During extended periods of withering, environmental factors such as temperature, humidity, and airflow significantly influence the physiochemical responses of the leaves, thereby causing variations in white tea quality (Tian et al., 2024). Regulating the rate of water loss is central to white tea production, as it orchestrates changes in the biochemical constituents of the leaves. Airflow is a key parameter influencing the rate of water loss and the progression of withering (An et al., 2022). Its management is essential for both traditional and mechanized processing methods (Huang et al., 2023). Nevertheless, comprehensive investigations into how airflow influences the withering process and its impact on white tea quality remain limited.

Air functions as a medium for heat and mass exchange during the withering process of white tea, with its velocity, volume, and direction significantly influencing dehumidification, water retention, and the oxygen supply. Inefficient air circulation can lead to the accumulation of undesirable gases, inhibiting polyphenol oxidation and disrupting the balance of leaf respiration (Yu et al., 2019). This imbalance prevents polyphenols from achieving an equilibrium between oxidation and reduction, resulting in the undesirable development of dull green or brown discoloration in white tea (Zhou et al., 2022). An increase in airflow in a withering environment accelerates water evaporation, prevents the accumulation of undesirable gases, and ensures an adequate oxygen supply to the withering leaves (Zhou et al., 2020). This process promotes polyphenol and enzyme activity, accelerates the accumulation of orthoquinones, and drives secondary oxidation reactions, contributing to the unique aroma, mellow taste, and refreshing characteristics of white tea. However, excessively high airflow velocity can shorten the withering duration, leading to incomplete hydrolysis and potentially grassy flavors in the tea (Feng et al., 2025). Therefore, a precise optimization of airflow parameters during the withering process is essential to ensure the desired quality of white tea. In production, installing air supply and dehumidification systems allows for precise control of environmental airflow, facilitating moisture evaporation from the leaf surface and promoting the transformation and accumulation of internal compounds (Qi et al., 2024). This ensures optimal withering and helps eliminate undesirable grassy odors, thereby enhancing the overall quality of white tea.

Airflow functions as a double-edged sword in the withering process of white tea, influencing its final quality by modulating the water loss rate and withering progression. To elucidate how airflow affects white tea withering and quality, we employed a synergistic combination of advanced analytical techniques, including low-field nuclear magnetic resonance (LF-NMR), transmission electron microscopy (TEM), scanning electron microscopy (SEM), gas chromatography-time-of-flight mass spectrometry (GC-TOF-MS), and ultra-performance liquid chromatography-triple quadrupole mass spectrometry (UPLC-TQS-MS). These techniques were utilized to monitor dynamic changes in leaf morphology, moisture content, microstructure, enzyme activity, and metabolite levels under varying airflow velocity conditions. The objective was to derive real-time feedback parameters from the withering process, thereby providing a reference for the intelligent judgment of white tea withering techniques. Additionally, our research aimed to identify the optimal airflow, offering foundational data for the operational parameters of withering equipment.

2. Materials and methods

2.1. Sample preparation and collection

The experimental materials consisted of freshly harvested tea shoots (Fuding Dabaicha cultivar, one bud with two to three leaves) from the tea garden of Fujian Agriculture and Forestry University in May 2021. The collected fresh leaves (FL) were allowed to wither in a room at a temperature of 24 ± 2 °C and a humidity of 63 ± 3 % (Fig. S1A and B). To create windless conditions, a portion of the FL was enclosed with gauze to surround the withering frames (Fig. 1A), resulting in an airflow velocity of 0 m/s (AF0). Another portion of the FL was subjected to different airflow velocities provided by fans, with airflow velocities of 0.5 m/s (AF0.5) and 1 m/s (AF1). The airflow velocity was measured via an anemometer (ZRQF30J, Minghe Zhike Technology Co., Ltd., Beijing, China). FL were withered for a total of 36 h, after which the leaves were dried to a moisture content of approximately 6 % via a roasting machine (JY-6CHX-70, Anxi Jiayou Machinery Co., Ltd., Fujian, China) at temperatures ranging from 90 to 100 °C, resulting in finished tea with airflow velocities of 0 m/s (F0), 0.5 m/s (F0.5), and 1 m/s (F1). During processing, samples were collected at different stages: FL, after 9 h, 18 h, 27 h, 36 h of withering, and from the finished tea. These samples were frozen in liquid nitrogen and stored at −80 °C for subsequent studies.

Fig. 1.

Characterization of white tea processing and leaf attributes under airflow regulation. (A) Schematic representation of the white tea processing workflow. (B) Variations in the leaf morphology and moisture content of white tea under different airflows during withering. (C) Color parameters (L*, a*, b*) of white tea processed under three different airflows. The data are presented as the means ± standard deviations (n = 3). Different letters indicate a significant difference at the P < 0.05 level.

2.2. Sensory evaluation and quantitative descriptive analysis (QDA)

The flavor profiles of different finished tea were assessed through QDA. A panel of six experts (three males and three females) with over 20 years of experience. Tea infusions were prepared according to the “Tea Sensory Evaluation Method” (National Standard of China, GB/T 23776-2018). Each 5 g tea sample was steeped in 150 mL of boiling water for 5 min, then filtered and presented to panelists in random order. The panel assessed the intensity of eight sensory attributes: “grassy,” “floral,” “pekoe,” “pure,” “fresh,” “mellow,” “astringent,” and “thickness” on a five-point scale (0 = none, 1 = recognizable, 2 = weak, 3 = medium, 4 = strong, and 5 = very strong). Mean scores for each attribute were used to generate radar charts.

2.3. Moisture content and color difference measurement

The moisture content of tea samples at different stages was measured using a fast moisture analyzer (SN-DHS, Lichen Technology Co., Ltd., Shanghai, China). Each sample was measured three times. The color of processing leaves and finished tea infusions was assessed using a portable colorimeter (YS3060, Lichen Technology Co., Ltd., Shanghai, China) in the color mode. Infusions were obtained during sensory evaluation. The colorimeter was calibrated with provided white and black panels. Three colorimetric indices: L*, a*, and b* were measured, indicating brightness (L*), red-green difference (a*), and yellow-blue difference (b*) respectively. Each color measurement was repeated three times.

2.4. LF-NMR relaxation measurements and magnetic resonance imaging (MRI) analysis

Tea samples were collected using a five-point method, with each sample (5 g) wrapped in plastic and analyzed in triplicate. Signals were acquired using an LF-NMR instrument (Niumag Analytical Instrument Co., Ltd., Suzhou, China) at 0.5 T (22 MHz, 40 mm coil). The magnet temperature was set at 25 °C. The spin-spin relaxation time (T2) was measured using the Carr-Purcell-Meiboom-Gill sequence (90-degree pulse length = 13.8 μs, 180-degree pulse length = 27.04 μs, echo time = 0.1 ms, wait time = 2000 ms, bandwidth = 200 kHz, echoes = 18,000, scans = 8). Relaxation data were analyzed using the SRIT regularization algorithm for multi-exponential fitting, generating relaxation spectra. These measurements were performed using the MultiExp Inv analysis software (Niumag Analytical Instrument Co., Ltd., Suzhou, China). MRI data were collected via a spin-echo sequence (echo time = 13.44 ms, repetition time = 600 ms, field of view = 120 mm × 120 mm, averages = 4, read size = 256, phase size = 192, slice width = 5 mm, slices = 1). Signal intensity measurements and analyses were conducted using the Osirix software (OsiriX Life v.7.0.4, Geneva, Switzerland).

2.5. Observation of tea cell structure and stomata

Three healthy mature second leaves from each tea sample were randomly selected, cut into 1 mm³ tissue blocks, and fixed in 2.5 % glutaraldehyde at 4 °C. TEM (Hitachi HT-7800, Tokyo, Japan) was used to observe the cell structure. After fixation, samples were rinsed thrice with 0.1 M phosphate buffer (pH 7.2), then fixed in 1 % osmium tetroxide for 2 h. After post-rinsing with buffer for 15 min and water washing for 10 min, the samples underwent gradual dehydration in an ethanol series (30 %–100 %) every 15 min, followed by three 20 min acetone dehydrations. Immersion in propylene oxide and embedding medium mixtures (1:1 and 1:2 at 37 °C and 45 °C, respectively) preceded polymerization at 60 °C. Sections were made with an ultramicrotome (Leica UC7, Hesse, Germany), double-stained with uranyl acetate and lead citrate, rinsed, and dried. The cell structure was observed and images were collected using TEM at 80 kV, with measurements taken for parameters related to cell walls, chloroplasts, and vacuoles in the images.

SEM (Hitachi TM3030, Tokyo, Japan) used to observe stomata on leaf tissue lower surface. Tissue removed from fixation solution, excess water wiped off, and tissue placed on sample stage with underside facing up. Stomatal images collected post-vacuuming, measurements taken for stomatal area (SAA) and stomatal density (SD), and the stomatal apertures area index (SAAI) calculated. The SAAI represents the stomatal aperture area per unit area of the leaf (Zheng et al., 2013). The calculation formula for SAAI is as follows:

$$\text{SAAI}=\mathrm{SD}\times \mathrm{SAA}\times 100\%$$ (1)

2.6. Enzyme activity determination

Enzyme activities of β-glucosidase, amylase, polyphenol oxidase (PPO), pectinase, peroxidase (POD), cellulase, lipoxygenase, and protease were determined using specific Elisa kits (Enzyme-Linked Biotechnology Co., Ltd., Shanghai, China). Fifty milligrams of fresh tea leaves were extracted for each assay. Activities were measured using a multifunctional enzyme-labeled instrument (MS-352, Labsystems Multiskan, Helsinki, Finland), with absorbance wavelengths set at 400 nm, 414 nm, 410 nm, 540 nm, 470 nm, 540 nm, 280 nm, and 366 nm for each respective enzyme. The specific detection process is shown in Fig. S2. Each sample underwent three independent repeat measurements.

2.7. Analysis of sugars, organic acids, and volatile compounds

2.7.1. Sample pretreatment

For volatile extraction, 2 g (freeze-dried) tea powder was placed in a headspace vial. A 1 cm, 65 μm polydimethylsiloxane-divinylbenzene (PDMS-DVB) headspace solid-phase microextraction (HS-SPME) fiber was incubated at 80 °C for 31 min, followed by 60 min of volatile collection. Volatiles were desorbed at 250 °C for 3.5 min, and analyzed using GC-TOF-MS. For sugars and organic acids, 30 mg (freeze-dried) tea powder was placed in a 2 mL centrifuge tube. The sample was treated with 480 μL of 75 % methanol and 20 μL of internal standard vanillic acid (5 mg/mL). The mixture was sonicated for 30 min, centrifuged at 12,000 rpm for 10 min, and 150 μL of the supernatant was vacuum dried. The sample was then treated with 80 μL of methoxyamine (20 mg/mL pyridine solution), heated at 80 °C for 20 min, and 80 μL of BSTFA +1 % TMCS was added. The sample was maintained at 70 °C for metabolite measurement. A quality control (QC) sample was prepared by mixing equal amounts of each sample, and a QC sample was collected after every three samples to monitor instrument stability. All samples underwent three biological replicates.

2.7.2. GC-TOF-MS analysis

Volatile compounds were detected using an Agilent 7890B gas chromatograph (Agilent Co., Santa Clara, CA, USA) coupled with a Pegasus HT time-of-flight mass spectrometer (LECO Co., Saint Joseph, MI, USA). Separation was achieved on a Restek Rxi-5Sil MS column (30 m × 0.25 mm × 0.25 μm) with helium as the carrier gas at 1 mL/min. The GC program commenced at 50 °C (5 min isothermal), ramping at 3 °C/min to 210 °C (3 min isothermal), with a transfer line temperature of 275 °C. The MS operated in electron impact (EI) mode (70 eV, 30–500 m/z) at 250 °C ion source temperature. For sugars and organic acids, the same column was used with a carrier gas flow rate of 1.5 mL/min. The GC program started at 80 °C (0.2 min isothermal), ramping at 15 °C/min to 160 °C, then 13 °C/min to 200 °C, and finally 10 °C/min to 300 °C (8 min isothermal). The transfer line temperature was 275 °C, and the MS operated in EI mode (−70 eV, 35–600 m/z) at 250 °C ion source temperature.

2.7.3. Qualitative and quantitative analysis of volatile compounds

Peaks of volatile compounds were identified by matching the National Institute of Standards and Technology (NIST) Mass Spectral Database. Retention indices (RI) were calculated using n-alkanes (C8–C30) and compared with database entries. Chemical structures, names, and CAS numbers were determined from PubChem (https://pubchem.ncbi.nlm.nih.gov/). Odor descriptions were sourced from the Flavor Ingredient Library (https://www.femaflavor.org/flavorlibrary/). The relative content of aroma components was determined by area normalization, calculating the ratio of each chromatographic peak area to the total peak area.

2.7.4. Calculation of relative odor activity value (rOAV) and aroma characterization influence value (ACI)

The rOAV and ACI were used to assess the contribution of aroma compounds to tea samples (Lin, Wu, et al., 2024). The rOAV of the component with the greatest aroma contribution was set to 100. Compounds with rOAV > 1 were considered key odor-active compounds significantly contributing to aroma characteristics. The odor characteristics and thresholds were obtained from literature (Hao et al., 2023; Zhu et al., 2024). The rOAV and ACI of each compound were calculated using the following formulas:

$${\text{rOAV}}_{\mathrm{i}}=\frac{{\mathrm{C}}_{\mathrm{i}}}{{\mathrm{C}}_{\max }}\times \frac{{\mathrm{T}}_{\max }}{{\mathrm{T}}_{\mathrm{i}}}\times 100$$ (2)

$$\mathrm{ACI}\left(\%\right)=\frac{{\text{rOAV}}_{\mathrm{i}}}{{\sum }_{\mathrm{k}}{\text{rOAV}}_{\mathrm{k}}}\times 100$$ (3)

In Formula (2), rOAV_i represents the relative odor activity value of each volatile compound; C_i represents the relative content (%) of each volatile component; C_max represents the relative content (%) of the compound with the greatest aroma contribution; T_i represents the threshold (μg/kg) of each volatile component; T_max represents the threshold (μg/kg) of the compound with the greatest aroma contribution. In Formula (3), ∑_krOAV_k represents the sum of the rOAVs of all volatile compounds.

2.8. Quantitative analysis of catechins, amino acids, and caffeine

2.8.1. Sample preparation

For amino acids and catechins, 30 mg of freeze-dried tea powder were extracted with 1 mL of 70 % methanol, vortexed, sonicated (30 min), and centrifuged (12,000 rpm, 10 min). The supernatant was filtered through a 0.22 μm membrane for UPLC-TQS-MS analysis. QC samples were prepared by mixing equal portions of each sample, with one collected for every 10 samples to monitor instrument stability. All samples were analyzed in triplicate.

2.8.2. UPLC-TQS-MS analysis

For catechin and amino acid analysis, 2 μL of the extract was diluted within the standard curve range and injected onto the Waters Acquity UPLC system coupled with the Waters photodiode array (PDA) detector and XEVO TQ-S MS triple quadrupole mass spectrometer (Waters, Milford, MA, USA) for metabolite measurement. Catechins were separated using a Waters Acquity UPLC BEH C18 column (2.1 mm × 100 mm, 1.7 μm) with a mobile phase of 0.1 % formic acid in water (A) and 0.1 % formic acid in acetonitrile (B). The gradient was: 95–83 % A (0–12 min), 83–0 % A (12–13 min), 0 % A (13–16.5 min), 0–95 % A (16.5–16.6 min), and 95 % A (16.6–20 min), at 0.3 mL/min. Amino acid were detected using a Merck SeQuant ZIC-HILIC column (2.1 mm × 100 mm, 5 μm) with a mobile phase of 5 mol/L ammonium acetate in water (A) and 0.1 % formic acid in acetonitrile (B). The gradient was: 5–41 % A (0–13 min), 41–60 % A (13–15 min), and 60–5 % A (15–20 min), at 0.4 mL/min.

2.9. Data processing and statistical analysis

Data were analyzed using Excel and SPSS 26.0 for one-way ANOVA to assess group differences (P < 0.05). Graphs were generated using GraphPad Prism 9.0 and Origin. Orthogonal partial least squares discriminant analysis (OPLS-DA) were conducted using SIMCA 14.1 software (Umetrics, Umea, Sweden). Heatmaps and hierarchical clustering were produced using TBtools (https://github.com/CJ-Chen/TBtools). Quantitative TEM and SEM image analyses were performed with ImageJ (National Institutes of Health, USA).

3. Results and discussion

3.1. Impact of airflow manipulation on the physical characteristics of white tea during withering

The morphological characteristics and moisture content of the leaves are critical indicators for assessing the progress of withering. The results indicated that freshly harvested tea leaves initially had a moisture content of approximately 78 % (Fig. 1B), characterized by fresh, upright leaves with a high moisture content. As withering progressed, significant changes in the physical properties of the leaves were observed under the three different airflows. Under the no-airflow treatment (AF0), water loss was slower, maintaining a bright green leaf color, and the moisture content remained significantly greater than that under the other airflow. After 36 h of withering, the moisture content dropped to 41 %, and the color of some leaves transitioned from deep green to blackish green. In contrast, water loss was faster under the airflow treatments AF0.5 and AF1. In the AF0.5 treatment, the leaves turned dark green at 27 h, and by 36 h, the moisture content had decreased to approximately 22 %. The leaves took on a gray-green hue, and some buds and leaves exhibited silver-white tones, which naturally shrank and curled. Under the AF1 treatment, water loss was the fastest, with leaves turning gray-green with reddish-brown tones by 27 h. By 36 h, the moisture content had decreased to 15 %, and the buds showed white pekoe. The leaf margins and backs curled naturally. Compared to AF0, the airflow-assisted withering treatments of AF0.5 and AF1 significantly accelerated water loss from the leaves and promoted the transformation of leaf color towards gray-green.

To accurately describe the color characteristics of white tea leaves during processing, a colorimeter was employed to measure the L*, a*, and b* values. Visual analysis revealed significant differences in color intensity and brightness among the three types of white tea (Fig. 1C). With increasing airflow intensity and withering duration, the brightness (L* value) of the leaves gradually decreased, whereas the a* value (representing the negative value of green) tended to increase, and the b* value (representing the positive value of yellow) progressively weakened. Throughout the withering process, as water evaporated from the leaves, it became wilted and lost luster. The increase in airflow intensity accelerated the loss of water and wilting of the leaves, leading to a continuous decline in the L* value. Additionally, withering caused a reduction in the chlorophyll a/b content, which is responsible for greenness, with an increase in the olive-colored phaeophytin a/b content (Herrera et al., 2022). Since the airflow treatments AF0.5 and AF1 accelerated the withering process, the degradation of chlorophyll and the formation rate of phaeophytin also increased, causing the leaf color to transition from green to gray-green and olive. As the process progressed, carotenoid compounds might also have undergone oxidative degradation, potentially causing the b* value of the leaves to gradually decrease (Ni et al., 2023). Furthermore, oxidative reactions during withering might have catalyzed the enzymatic oxidation of catechins and other substances, leading to the formation of pigment substances like theaflavins and the reddish-brown color observed in the withering leaves of AF1 (Long et al., 2024). These results indicate that among the three groups of samples, the moderate airflow velocity of AF0.5 was more conducive to the color transformation of white tea.

3.2. Analysis of moisture phase transitions and migration pathways affected by airflow during the withering process of white tea

Withering is the core process in white tea processing and fundamentally involves dehydration stress to trigger physiological and chemical reactions within the leaves. The essence of withering under different airflow lies in the regulation of the water loss rate and the withering progress. The results indicated that the three white tea treatments resulted in significant differences in moisture content and water composition. To explore moisture dynamic changes during processing, LF-NMR and MRI were used to analyze samples processed under three different airflows (Fig. 2A). LF-NMR, a non-invasive moisture detection method, uses the T2 of the inversion spectrum to show the chemical environment of hydrogen protons in the sample, reflecting moisture phase state, fluidity and migration (Cui et al., 2019). The T2 relaxation times ranged from 0.01 to 10,000 ms, categorizing the internal moisture of the tea leaves into three states: free water, immobilized water, and bound water. Specifically, 0.01–1 ms relaxation times mainly correspond to bound water tightly linked with cell wall fibers etc. 1–10 ms represents immobilized water in the cellular plasma, which is present mainly in the forms of enzymes, hydrated hemicelluloses, intermediate metabolites, and macromolecules. 10–1000 ms correspond to free water in a free state, primarily vacuolar water, with high mobility (Cheng et al., 2019). T2 and signal amplitude curve analyses demonstrated that as the withering process progressed, the water content in the leaves gradually decreased, leading to significant leftward shifts in the signal amplitude curve due to changes in the internal structure of the samples (Fig. 2B). Compared with those of the no-airflow treatment AF0, the curves of the airflow treatments AF0.5 and AF1 shifted more to the left, indicating substantial structural changes within the leaves and reduced binding of hydrogen protons.

Fig. 2.

Nondestructive analysis of moisture dynamics in white tea processing under varying airflows via LF-NMR and MRI. (A) Schematic diagram of nondestructive detection via LF-NMR and MRI systems. (B) Transverse relaxation curves of white tea processed under different airflows. (C) Proportional distribution of water component structures in white tea processing. (D) MRI images of three distinct processing stages.

The relative content of moisture phases can be reflected from the corresponding peak areas in the T2 relaxation spectrum (Fig. 2C). By converting these peak areas into the proportions of each phase state of moisture in the leaves, we found that FL contained a high level of free water, which dominated the composition (Table S1). As withering proceeded, the proportion of free water significantly decreased, with a large amount of free water lost from the leaves. Before 18 h of withering, the proportions of bound water and immobilized water in AF0.5 and AF1 increased with increasing airflow intensity, indicating that airflow promoted the loss of free water in the early stage of wilting and facilitated the transformation of moisture into an immobilized and bound state. By 27 h, AF0.5 had the highest proportion of immobilized and bound water, whereas AF0 had the highest proportion of free water. Compared to the no-airflow treatment AF0, samples subjected to airflow-assisted withering presented a faster rate of water loss, a higher proportion of bound water, and moderate airflow velocity (AF0.5) were found to facilitate the conversion of other forms of water into bound water. By 36 h of withering, the proportion of bound water increased with increasing airflow intensity. At this stage, the AF0 samples still contained a high amount of free water, but AF1 had the highest proportion of bound and immobilized water, and AF0.5 had the highest proportion of free water. The results indicated that the latter stage of withering is characterized by a rapid decline in moisture content, with other water components gradually converting to bound water. After withering, all three samples were heated at 90–100 °C until the moisture content reached approximately 6 %, at which point the moisture content drastically decreased. Heating caused the tea leaf structure to shrink, resulting in tighter bonding between semi-bound water and macromolecules. Compared with the other samples, the finished tea F1 had the highest content of bound water, whereas F0.5 had higher contents of immobilized and free water than other samples. This suggests that airflow-assisted withering promotes the transformation of other phase state moisture into bound water and that moderate airflow significantly increases the content of immobilized and free water.

MRI is an technique used for visualizing the spatial distribution and structural changes of water inside plants. The higher the water content of a plant, the stronger its image signal, allowing for the analysis of water migration paths through changes in the image signal (Ezeanaka et al., 2019). By examining proton density images of leaves at different stages of withering, the uneven distribution of water in the leaves was observed (Fig. 2D). FL displayed a complete leaf outline, with clear veins and leaf margins. The signal intensity was predominantly concentrated on the stems and buds, indicating a fresh and vibrant state with an abundant moisture content in the stems, leaves, and buds (Fig. 2D-a, b, c). Before 18 h of withering, the signal intensity of the stems and buds in all three groups of withering leaves decreased. The leaves of AF0 were in a stage of balanced water loss, with the signal at the leaf tips starting to blur at 18 h (Fig. 2D-d, g). The leaf outline of AF0.5 gradually blurred, although the vein boundaries remained relatively clear, and the signal intensity of the buds significantly weakened (Fig. 2D-e, h). The leaves of AF1 showed unclear boundaries after 9 h, whereas the buds exhibited strong signals (Fig. 2D-f). This finding indicates that during the early stages of withering, the leaves in the no-airflow treatment experienced balanced water loss, whereas the airflow treatment accelerated water loss from the leaves. During the withering process, water migrates from high-signal areas to low-signal areas, with different migration paths under different airflows. The moderate airflow velocity (AF0.5) allowed water from the stems to be preferentially transported to the leaves, making the veins and leaf outlines more pronounced (Fig. 2D-b, e, h). In contrast, the leaves of AF1 experienced significant water loss, with water transferred from the stems to the buds (Fig. 2D-c, f, i). The differences in the image information of the leaves in the late withering stage were more pronounced. Leaves subjected to no-airflow treatment had gradually blurred boundaries, with stems still retaining a high water content (Fig. 2D-j, m). The leaves subjected to airflow-assisted withering presented a significant decrease in moisture content, with the leaf outlines almost invisible, leaving only a portion of the stems with a relatively higher moisture content (Fig. 2D-k, n, l, o).

3.3. Dynamic influence of airflow on the withering process of white tea: A comprehensive analysis from phenotypic characteristics to cellular structures

During the withering stage, white tea leaves subjected to different airflow treatments experienced varying degrees of water loss stress, resulting in tissue damage and triggering the release of stress signaling molecules, which in turn initiated a series of physiochemical reactions. When discussing phenotypic plasticity, it is essential to consider the hierarchical structure of the tissues. To better understand the underlying causes of the physiochemical changes in the leaves, SEM and TEM were employed to observe the microstructural changes in the leaf cells during processing (Fig. 3A).

Fig. 3.

Effects of withering under different airflow on the leaf phenotypes and microscopic structures of white tea. (A) Variations in the leaf phenotypes and microscopic structures of white tea during withering under different airflows. (B) Quantification of parameters including the SD, SAAI, cell wall thickness, and vacuole area ratio, in stomatal and cellular structures. The data are presented as the means ± standard deviations (n = 3). Different letters indicate a significant difference at the P < 0.05 level.

Freshly picked tea leaves had a moisture content of approximately 78 %, characterized by vibrant green and upright, plump buds and leaves (Fig. 1B). At this stage, the leaf tissue contained abundant immobilized and free water (Fig. 2C), which was present in a detached state. This process resulted in the closing of a small number of stomata to maintain the water balance of the leaves (Fig. S3). The stress stimuli triggered coordinated responses within different organizational levels of the plant. At the cellular level, the cells swelled due to water saturation, with small intercellular spaces and intact structures. The organelles were neatly arranged along the cell membrane (Fig. S4). These microstructural changes helped the leaves maintain a fresh and upright morphology.

After 18 h of withering, water loss stress caused the moisture content of the leaves under the different airflow treatments to decrease to between 43 % and 64 %, and the tea shoots exhibited bending and wilting characteristics. With increasing airflow intensity, a significant amount of free water in the leaves was lost, promoting the transformation of water into immobilized and bound states. During this stage, the stomata adjusted water and gas exchange by increasing the number and area of apertures. AF0 exhibited the lowest stomatal density and the greatest number of closed stomata (Fig. 3B). Compared with AF1, AF0.5 exhibited the greatest stomatal density but had smaller aperture areas and elongated elliptical shapes. The cellular responses also differed under various treatments. The cells in AF0 showed a more regular morphology, with thinning cell walls and membranes and some damaged chloroplasts and vacuoles (Fig. S5). As the airflow velocity increased, the withering progression intensified, causing the cells to exhibit slight vacuolation and an uneven distribution of substances. The degree of cell deformation increased, which was evident in significant plasmolysis and widening of the intercellular spaces. The cell walls exhibited uneven chain-like structures, with damaged chloroplasts and vacuoles, and most of the free water in the vacuoles evaporated, causing shrinkage. These physiological and structural changes caused the leaves to shrink and wither due to water loss, with increased airflow exacerbating the shrinkage of the leaves and buds.

After 36 h of withering, prolonged water stress caused the leaf moisture content to decrease to 15–41 %, reaching peak levels of bound water. Concurrently, the leaves transitioned to a dark green hue, with the buds and leaves exhibiting signs of shrinkage. At this stage, the leaves exhibited new phenotypic characteristics at multiple scales due to varying degrees of water loss signaling. The leaves of AF0 turned dark green, with black-brown patches at the tips. In contrast, the leaves of AF0.5 and AF1, which were subjected to airflow-assisted withering, turned from dark green to grayish green, and the buds appeared silvery white color. Compared with that in AF0.5, the leaf color distribution in AF1 was more uneven. Compared with those of the 18 h stage, the stomatal density and aperture area of the AF0 stage increased. The increase in airflow resulted in a continuous increase in stomatal density and aperture area for AF1. The AF0.5 presented the lowest stomatal density and aperture area, with most stomata gradually closing. At this stage, the loss of water caused the stomata to open and close to varying degrees, leading to curled leaf margins and twisted buds and stems. At the cellular level, due to damage and water loss, the cells collapsed, and the internal cell structures became blurred due to the leakage of contents. With the intensification of airflow-assisted withering, chloroplast structures gradually deteriorated, and the cell walls exhibited chain-like fractures due to lysis (Fig. 3A). The complete rupture of central vacuoles led to the massive overflow of free water, polyphenols, and enzymes, with the substrates, water, and large sugar molecules bound in the cytoplasm reacting with the substances flowing out of the vacuoles, consequently causing the conversion of free water into bound water.

3.4. Effects of airflow on enzymatic activity and chemical reactions in white tea during withering

During prolonged withering of white tea, the continuous loss of leaf moisture enhanced respiratory activity, leading to increased cell sap concentration and cell membrane permeability. This triggered physiochemical changes in intercellular proteins, causing oxidases, hydrolases, and other enzymes to transition from bound states to free states, significantly enhancing the activity (Wang et al., 2024). The increased enzyme activity subsequently catalyzed the oxidation and hydrolysis of macromolecular compounds, generating small, simple compounds that lay the foundation for the formation of tea color, aroma, and flavor substances (Zou et al., 2024). To further investigate the chemical reactions occurring in white tea during the withering stage with different airflows, the activities of relevant enzymes were measured. PPO and POD are crucial catalytic enzymes involved in the oxidation of polyphenols to tea pigments. During the withering process, PPO activity in leaves continuously decreased under AF0, whereas airflow-assisted withering initially decreased PPO activity before it subsequently increased (Fig. 4A). PPO oxidizes polyphenols to form phenolic quinones, and the accumulation of phenolic quinones exerts feedback inhibition on PPO, leading to a decrease in enzyme activity (Guo et al., 2021). However, as cell dehydration intensifies, PPO activity also increases. Under the AF0.5 treatment, the enzyme activity of POD tended to decrease, whereas under the AF0 and AF1 treatments, the enzyme activity initially increased before subsequently decreasing (Fig. 4B). The gradual loss of water during POD activity led to autolysis of the enzyme protein, causing decreased activity. Moreover, PPO oxidizes catechins into theaflavins, which are then catalyzed by POD to form thearubigins (Zhou et al., 2023). At 36 h of withering, POD exhibited the highest activity under the AF1 treatment, which may explain the reddish-brown coloration of the leaves. Additionally, a series of consecutive hydrolytic reactions occurred during the withering period. Amylase activity tended to increase in the AF0.5 treatment group but initially declined before rising in the AF1 treatment group. In contrast, amylase activity consistently decreased in the AF0 treatment (Fig. 4C). As one of the macromolecules of sugars, starch is considered an important molecule that mediates the plant response to a water deficit (Thalmann & Santelia, 2017). The loss of water during the withering process triggered amylase activity, leading to starch degradation into sugar substances. This process indicated that the degree of water deficit and amylase activity were not linearly related. Under moderate airflow velocity (AF0.5) treatment, amylase activity in the leaves was the highest. With respect to protease activity, a significant decrease was observed under the AF0 treatment, whereas a notable increase occurred under the AF0.5 treatment. Slightly reduced activity was observed under the AF1 treatment (Fig. S6A). Proteases are enzymes that catalyze the hydrolysis of proteins, resulting in the production of amino acids and other substances that enhance the fresh taste and aroma of the tea (Hu et al., 2024).

Fig. 4.

Changes in enzyme activity during the withering process of three groups of white tea. Variations in the activities of PPO (A), POD (B), amylase (C), cellulase (D), β-glucosidase (E), and lipoxygenase (F) during the withering stage of white tea under different airflow treatments. The data are presented as the means ± standard deviations (n = 3). Different letters indicate a significant difference at the P < 0.05 level.

The cell wall is the primary structure that maintains cell morphology and regulates osmotic pressure in tea leaves and is composed of polysaccharides such as cellulose, hemicellulose, and pectin (Bidhendi et al., 2020). Continuous water loss in leaves triggered cell deformation and structural damage, subsequently enhancing interactions between intracellular substrates and enzymes. These changes were closely related to the activities of cellulase and pectinase. Under different airflow treatments, enzyme activity exhibited distinct dynamic changes, and the increase in these enzyme activities further accelerated cell wall degradation. In the AF0 treatment, cellulase activity initially decreased but then increased (Fig. 4D). In contrast, under the AF0.5 and AF1 treatments, cellulase activity initially increased but then decreased, with slightly greater cellulase activity at 18 h observed in the AF0.5 treatment compared to the AF1 treatment. The hydrolysis reaction between cellulose and cellulase released more substrates for PPO and POD (Hu et al., 2024). Pectinase activity initially increased but then decreased under both the AF0 and AF1 treatments. At 18 h of withering, pectinase activity reached its peak under the AF0 and AF1 treatments and then decreased. In contrast, pectinase activity continued to increase under the AF0.5 treatment (Fig. S6B). Pectin is hydrolyzed under the action of pectinase into soluble sugars and pectin, increasing its cellular viscosity, and pectinase plays a crucial role in cell wall degradation (Du et al., 2022). Additionally, β-glucosidase also contributes to cell wall degradation. Its activity initially increased but then decreased under all three airflow treatments, with the highest activity observed under AF0 and the lowest under AF0.5 (Fig. 4E). β-Glucosidase undergoes hydrolysis reactions with glycosides stored in vacuoles, promoting the release of terpenoids, which influences the formation of tea aroma (Chen et al., 2019). Lipoxygenase, as an enzyme involved in the stress response, exhibits different dynamic characteristics in withering white tea leaves under various airflow treatments (Zou et al., 2024). Under AF0, lipoxygenase activity consistently declined, whereas under AF1, it initially decreased before slightly increasing (Fig. 4F). Under AF0.5, the enzyme activity initially increased but then decreased, indicating that enzymatic reactions occurred primarily during the early withering stages. Lipoxygenase catalyzes the oxidation of fatty acids such as α-linolenic and linoleic acids, generating a series of volatile compounds derived from fatty acids, such as alcohols and aldehydes, which contribute to the aroma of tea (Ho et al., 2015).

3.5. Airflow-mediated alterations in aroma compounds and metabolites in white tea during withering

Different airflow induced varying degrees of enzymatic reactions in white tea during withering due to dehydration stress, subsequently affecting the synthesis and transformation of aroma compounds. To explore the changes in aroma components in the three types of white tea with different withering process in depth, a comprehensive analysis of volatile compounds was conducted using GC-TOF-MS. A total of 101 volatile components, including 26 alcohols, 20 esters, 17 aldehydes, 12 ketones, 6 acids, 13 hydrocarbons, and 7 other compounds, were screened (Table S2). The analysis revealed that alcohol compounds were the primary components constituting the volatile substances of white tea (Fig. S7), capable of combining with sugars to form glycosides stored in the tea leaves (Chen et al., 2022). During early withering (18 h), alcohols content increased with airflow, whereas the esters content decreased. At this stage, the samples subjected to airflow had higher levels of ketones and acids, whereas the no-airflow withering samples gained more esters and hydrocarbons. This finding indicated that cell damage was caused by increasing airflow during early withering, facilitating the conversion of alcohols to esters. By 36 h of withering, the total aroma content of AF1 drastically decreased. The moderate airflow velocity AF0.5 sample had higher contents of esters, ketones, and hydrocarbons than other treatments. The samples were effectively separated by withering type using OPLS-DA (Fig. 5A). Notably, the volatile compounds presented only minimal differences at 36 h compared with those at 18 h under AF0, suggesting relatively stable changes in this stage. Combined with the results of 200 permutation tests, the model performed well in terms of fitting degree R²Y (cum) = 0.988 and predictive ability Q² (cum) = 0.933, without signs of overfitting (Fig. 5B).

Fig. 5.

Variations in aroma compounds and metabolites of white tea during withering under different airflow. (A) OPLS-DA score plot. (B) OPLS-DA model cross-validation plot. (C) Heatmap and VIP scatter plot of differential volatile compounds. (D) Odor description wheel for key differential compounds. (E) Heatmap of white tea metabolite changes. The analysis is based on the normalized average signal abundance from three biological replicates per sample.

To delve deeper into the key aroma components during the withering process, a variable importance projection (VIP) plot was constructed, where volatile substances with VIP > 1 were identified as important components distinguishing aroma characteristics (Wang et al., 2020). Combined with the criteria of VIP > 1 and P < 0.05, 21 volatile compounds with differences were identified (Fig. 5C), which could serve as potential markers to differentiate the volatile components of samples at different stages. However, GC-TOF-MS results alone do not reflect compound contributions to aroma, which depends on odor thresholds (OT) and contents. Thus, the rOAV and ACI were calculated for the 21 compounds to assess the aroma contributions. Compounds with rOAVs ≥ 1 were deemed key flavor contributors (Lin, Wu, et al., 2024). The top five ACI were also calculated to further evaluate the aroma contribution of the samples. Analysis revealed that linalool was the dominant aroma contributor among the seven screened compounds, with 14 substances having an rOAV ≥ 1, showcasing fresh and green odors (Table S3). In FL, 12 key contributors were identified, with linalool, nerol, linalool oxide II, nonanal, and (E)-jasmone exhibiting the highest ACI, contributing 91.98 % to the overall aroma. Additionally, phenylethyl alcohol, indole, (E)-jasmone, and methyl salicylate were identified as key differential substances in FL, primarily exhibiting floral and green scents (Fig. 5D).

At 18 h of withering, the five compounds with the highest ACI in the AF0 sample contributed 90.65 % of the overall aroma. Specifically, (Z)-3-hexenyl hexanoate (fruity green odor) and nonanal (floral, green) had ACI greater than those of the airflow-assisted withering samples. The key differential substances in AF0 at this stage, which primarily contribute to fruity and green odors, included linalool oxide II, (Z)-3-hexenyl hexanoate, (E)-2-hexen-1-ol, linalool oxide I, and methyl salicylate. For the airflow-assisted withering sample AF0.5, the five compounds with the highest ACI contributed 92.84 % of the aroma. Linalool, which is rich in floral scents, had a higher ACI than the other samples. Phenylethyl alcohol, 3-hexen-1-ol, linalool, and (Z)-3-hexenyl hexanoate were identified as key differential substances in AF0.5. In AF1, nerol (floral, fruity) and 1-hexanol (green, herbal) had higher ACI, contributing to a floral and green aroma profile. The key differential substances in AF1 included phenylethyl alcohol, 3-hexen-1-ol, nerol, and linalool oxide II, which collectively exhibit a floral and green aroma. These results indicated that regulating airflow velocity to control the withering process resulted in differences in key differential substances between the no-airflow and airflow-assisted withering samples, thus resulting in different aroma types. The aroma profiles of the airflow-assisted withering samples AF0.5 and AF1 were similar. Notably, no-airflow withering promoted the accumulation of green scents, as evidenced by the presence of nonanal in AF0. At 36 h of withering, the aroma differences among the samples became more pronounced. The top five compounds in AF0, AF0.5, and AF1 contributed 94.51 %, 93.31 %, and 92.1 % to the overall aroma, respectively. Nerol (floral) and nonanal had higher ACI in AF0 than in the airflow-assisted samples. Linalool had the highest ACI in AF0.5, indicating that it made the greatest contribution to the aroma. In AF1, linalool oxide II (floral), (Z)-3-hexenyl hexanoate, and 1-hexanol exhibited higher ACI. The key differential substances in AF0 included linalool, nerol, linalool oxide II, 1-hexanol, (E)-2-hexen-1-ol, linalool oxide I, and nonanal, which collectively produce floral and green aromas. AF0.5 had phenylethyl alcohol, linalool, 1-hexanol, and (Z)-3-hexenyl isovalerate as key substances, primarily showcasing floral and fruity odors. The different substances in AF1 included 1-hexanol, linalool oxide I, and (Z)-3-hexenyl isovalerate, which exhibit floral, fruity, and green odors overall. Compared with AF1, AF0.5 had richer key different substances and more pronounced floral expression. These findings suggest that excessive or insufficient water loss during the withering process of white tea might promote the further accumulation of green odor substances, whereas moderate water loss could facilitate the development of a floral aroma.

During the tea processing, aroma formation of aroma involves the biosynthesis of various metabolites, such as fatty acids, amino acids, and carotenoids, which ultimately transform into volatile compounds (Ho et al., 2015). This complex metabolic network not only provides the foundation for aroma formation but also lays the groundwork for taste attributes, collectively shaping the unique flavor of tea. Amino acids are the main compounds determining the flavor of white tea. As withering progressed, amino acids in the AF0.5–36 samples were significantly enriched, with the concentrations sharply increasing (Fig. 5E). The increase in protease activity under AF0.5 conditions promoted protein hydrolysis and amino acid release, as evidenced by the sharp increase in amino acid concentration. During the later stages of withering, polyphenol oxidation generated ortho-quinones that reacted with amino acids to form aldehydes and esters, contributing to the aroma of white tea (Chen et al., 2019). Simultaneously, the caffeine content gradually decreased with increasing withering process and airflow, which helped reduce the bitterness of white tea. Catechins, the main polyphenol compounds in tea, have significant astringent and bitter flavors. In the early stages of withering, the contents of ester-type catechins (ETC) and non-ester-type catechins (NETC) in AF0 were significantly greater than those in the airflow-assisted withering samples and gradually decreased with increasing airflow. The accumulation of GCG and EGCG-CH3 in airflow-assisted withering samples could be attributed to the gradual acidification of the internal tissues due to cell damage, promoting the isomerization of EGCG to GCG (Wang et al., 2023). In the later stages of withering, catechins in the leaves undergo oxidation reactions catalyzed by PPO and POD, converting to quinones and catechin oxidation products, leading to a downward trend in overall catechin levels.

During dehydration stress in withering leaves, phenotypic changes lead to the transformation of water and soluble substances, with sugars acting as signaling molecules for intercellular transport (Yue et al., 2015). Among the different stages of the samples, we identified nine sugar substances, with FL having high levels of arabinose and glucose-1-phosphate. In the early stages of withering, most sugar levels decreased, indicating that dehydration stress accelerated the respiratory process, increasing the consumption of sugars as respiratory substrates (Lin, Huang, et al., 2024). In the later stages of withering, the airflow-assisted samples exhibited significantly greater sugar contents than the no-airflow samples, likely because the increased airflow velocity caused cell damage and promoted the enzyme-catalyzed hydrolysis of cellulose and starch, thus increasing soluble sugar levels (Zhou et al., 2022). Organic acids, which are abundant in tea, provide structural integrity and energy for metabolic processes and play a role in signal transduction (Fan et al., 2021). During withering, 12 organic acids including stearic acid, α-linolenic acid, linoleic acid, and elaidic acid, were detected. The FL contained high levels of lactic acid, malonic acid, and malic acid. At 18 h of withering, the leaves under the AF0.5 treatment exhibited greater total organic acid (TOA) and oxalic acid levels compared to those under the other treatments. Leaves treated with AF0 and AF1 exhibited relatively higher levels of shikimic acid, a metabolite involved in the shikimate pathway leading to the biosynthesis of aromatic amino acids (Zou et al., 2022). At 36 h of withering, the leaves under the AF0 treatment showed higher levels of TOA and oxalic acid than the airflow-assisted samples. In AF0.5, stearic acid, α-linolenic acid, linoleic acid, and elaidic acid were more abundant, serving as key precursors for volatile fatty acid derivatives that can be easily degraded into floral and green-scented volatiles under enzyme action (Yang et al., 2013). Lipoxygenase, a key enzyme in the metabolism of α-linolenic acid and linoleic acid, promoted the conversion of unsaturated fatty acids into alcohols and aldehydes (Wu et al., 2022).

3.6. Airflow manipulation during the withering process affects the quality and flavor profile of white tea: Insights from color, aroma, taste, and overall sensory evaluation

Different processing methods significantly affect the color, aroma, and taste of the final tea product due to dynamic interactions among physiological, chemical, and metabolic changes during processing. Withering experiments with varying airflows on tea leaves yielded three finished tea products (F0, F0.5, and F1), which were evaluated for sensory quality. The F0 samples had a green color, whereas the airflow-assisted withering samples (F0.5 and F1) had a grayish-green color, indicating that airflow-assisted withering accelerated chlorophyll conversion, achieving the desired color quality for normal white tea (Table S4). Significant differences in tea infusion color were observed among the three samples. Combined with the analysis of L*, a*, and b* values, indicated that moderate airflow withering brightened the infusion color of F0.5, while no-airflow withering darkened the infusion color of F0 (Fig. 6A). As airflow increased, the a* and b* values in the tea infusion gradually increased. F0 had a greenish-yellow color due to low chlorophyll conversion, whereas F0.5 and F1 had an apricot-yellow color from polyphenol oxidation into pigments such as thearubigins and theaflavins. These soluble pigments, combined with other internal components, formed the apricot-yellow infusion and grayish-green appearance. We subsequently used five descriptive terms to evaluate the aroma of the samples through olfactory assessment (Fig. 6B). F0 exhibited a pure aroma with slightly floral and grassy notes. F0.5 was characterized by a fresh, pekoe, and floral scent, whereas F1 displayed a more pronounced grassy aroma. Moderate airflow velocity withering significantly enhanced the floral attributes of white tea. By tasting the tea infusions, we described their taste attributes (Fig. 6C). The results showed that the tea infusion of F0 exhibited strong astringency. F0.5 offered a mellow and thick taste with a hint of freshness. In contrast, F1 showed a stronger grassy aroma with reduced mellowness and thickness. No-airflow withering promoted the formation of astringent substances, whereas moderate airflow withering facilitated the development of normal taste qualities in white tea. Excessive airflow withering, due to rapid water loss and rapid content conversion, results in a grassy taste in the tea infusion.

Fig. 6.

Differences in color, aroma, and taste quality of finished tea. (A) Colorimetric parameters of tea infusion (L*, a*, b*). (B) QDA radar chart depicting the aroma characteristics of tea. (C) QDA radar chart illustrating the taste attributes of tea leaves. (D) OPLS-DA score plots of three types of finished tea. (E) Heatmap and VIP scatter plot of differential volatile compounds in finished tea. (F) Key differential compounds and their characteristic aroma attributes. (G) Heatmap illustrating changes in taste compounds of finished tea. The analysis is based on the normalized average signal abundance from three biological replicates per sample.

Aroma is one of the essential attributes of tea quality, with different processing environments impart diverse aroma types to the tea products. After drying, we observed that the relative content of volatile substances in the finished tea gradually increased with increasing airflow (Fig. S8A). Specifically, the relative contents of esters, aldehydes, ketones, and other substances all increased with airflow. The F0.5 samples were rich in alcohol substances, whereas no airflow conditions were conducive to the accumulation of acids and hydrocarbons. Significant differences in volatile components were present among the three tea products, which were distinguishable in the OPLS-DA score plot (Fig. 6D). The reliability of the model was confirmed by the test results (R²Y = 0.998, Q² = 0.981) (Fig. S8B). Based on the screening criteria of VIP > 1 and P < 0.05, we identified 17 differential volatile substances (Fig. 6E). These substances are distributed in regions I, II, and III and play important roles in tea aroma. Based on the criterion of an rOAV > 1, 9 key substances that exhibit different aroma characteristics through synergistic effects were identified (Table S5). Linalool had the highest rOAV, with strong floral notes, significantly contributing to the aroma. With increasing airflow, the ACI of linalool gradually increased, indicating a significant increase in its contribution to the aroma of the samples. Linalool, derived from sugar derivatives via β-glucosidase hydrolysis, serves as the primary substance base for the pekoe and floral qualities of white tea (Ho et al., 2015). In the F0 sample, (E)-2-hexen-1-ol and 1-hexanol were identified as key differential substances, mainly presenting green and fruity aromas. These substances originate from fatty acid oxidation, resulting in a green aroma in the finished tea. In the F0.5 sample, nerol, linalool, linalool oxide I, and linalool oxide II were key differential substances, resulting in floral, fruity, and sweet aromas. These substances are all alcohols, primarily related to linalool and its oxides. Linalool possesses chemically unstable properties, and its accumulation occurs during intensified leaf desiccation and cellular damage, reflecting the active metabolism of alcohol substances within cells. In combination with linalool, linalool oxide II, phenylethyl alcohol, linalool oxide I, (E)-5,6-epoxy-β-ionone, and 3-hexen-1-ol, the F1 sample exhibited aromas of green, floral, fruity, and woody notes. Although the key differential substances in the F1 sample were similar to those in the F0.5 sample, the aroma of F1 was richer, with a distinct green scent and a blend of other aromas. These substances are mostly glycosidase hydrolysis products, such as linalool and its oxides, phenylethyl alcohol, etc. Notably, (E)-5,6-epoxy-β-ionone is a carotenoid monoterpene compound with an epoxy group at positions 5 and 6, offering a sweet berry and woody aroma. Its generation might be due to complex chemical reactions occurring in an oxidative environment under water stress, leading to the oxidation of β-ionone (Zou et al., 2024). Different airflows altered the water loss rates and biosynthetic pathways of tea leaves, resulting in different volatile compounds and distinct aroma characteristics in finished tea.

The taste of tea is a comprehensive sensory experience resulting from the interaction of water-soluble substances in the tea infusion with human taste receptors. The main components contributing to tea taste include organic acids, tea polyphenols, caffeine, amino acids, and soluble sugars. Amino acids confer a refreshing taste and undergo reactions such as Strecker degradation and Maillard during roasting, converting into aromatic compounds in high-temperature environments (Wu et al., 2020). The F0 samples exhibited the lowest total amino acid (TAA) content but were enriched in Asp and Tyr. Asp, imparting an umami taste, followed a trend of first increasing but then decreasing during white tea withering (Dai et al., 2017). In this study, the lack of airflow subjected the F0 sample to a slow withering process, which likely contributed to the enrichment of Asp in the F0 sample. Simultaneously, the content of Gly increased with increasing airflow. The F0.5 sample had significantly higher levels of other amino acids and TAA, enhancing the refreshing taste of its infusion. Caffeine, the main source of bitterness, decreased with increasing airflow, potentially resulting in a mellower taste of the tea infusion. Caffeine can form complexes with theaflavins through hydrogen bonds, increasing the refreshing taste, which is closely related to polyphenol and tea pigment conversion. Catechins, as the primary components of polyphenols, are typically recognized as the key contributors to the astringent taste of tea (Zhuang et al., 2020). As airflow increased, the TC content gradually decreased, indicating that airflow mitigated the astringency of the tea infusion. In high-temperature environments, catechins undergo pyrolysis reactions, leading to synthesis and isomerization reactions, thereby generating substances such as tea pigments. Sugars are the primary source of sweetness in tea infusions. The F1 samples exhibited the highest TS content, indicating that an airflow of 1 m/s facilitated the accumulation of sugar substances, thereby positively contributing to the sweetness of the tea infusion. Conversely, the F0.5 sample had the lowest TS content, possibly due to the condensation reactions between amino acids and sugars during the drying stage, which formed aroma compounds with sweet notes (Qu et al., 2019). Some sugars undergo caramelization reactions, generating dehydrated polymers like caramel, further consuming sugar substances. Organic acids are important components that influence the acidity and fullness of the tea infusions. Moderate amounts of acidic compounds can enhance the fullness of the taste (Shirai, 2022). Compared with the other samples, the F0.5 sample exhibited significantly higher TOA and elaidic acid contents, whereas the TOA contents of both the F0 and F1 samples were relatively low. α-Linolenic acid and linoleic acid were enriched in F0.5 and F1, with α-linolenic acid generating volatile compounds through the lipoxygenase pathway, laying the foundation for the formation of tea flavor (Zou et al., 2024). Tea taste is achieved through the superposition, interaction, synergy, and inhibition of various components, ultimately balancing the flavor and taste in tea products.

In conclusion, as a pivotal regulatory factor in white tea withering, wind speed profoundly influenced the physicochemical properties and sensory quality through dynamic modulation of water migration pathways, cellular structural integrity, and enzymatic reaction networks. Under windless conditions, water retention led to the accumulation of green and astringent compounds (e.g., (E)-2-Hexen-1-ol and catechins). Although a wind speed of 1 m/s accelerated the withering process, excessive water loss caused the presence of grassy odors (e.g., 3-hexen-1-ol) and insufficient hydrolysis. In contrast, a moderate wind speed of 0.5 m/s precisely balanced the water loss rate and the progress of physicochemical reactions. It not only promoted the formation of bound water, the transformation of the tea color to grayish-green, and the accumulation of floral aroma substances (such as linalool and nerol), but also released enzymes and substrates through controlled cell damage, facilitating the enrichment of amino acids and organic acids. Ultimately, it resulted in a mellow, fresh, and refreshing flavor profile.

The study established that wind speed served as the principal determinant in coordinating oxidative, hydrolytic, and biosynthetic pathways during the withering of white tea. Optimizing wind speed could effectively address the issue of unstable quality in traditional processing. These findings provided a theoretical basis for setting wind speed parameters and designing withering equipment in white tea production. Wind speed was the “invisible hand” that regulated the quality of white tea during withering. Only by scientifically controlling its intensity could we achieve a delicate balance between water-loss stress and biochemical reactions, creating the unique and harmonious quality of white tea in terms of color, aroma, and taste.

4. Conclusion

This study employed a multi-layered approach to systematically investigate the impact of various airflow on the withering process of white tea, with a focus on the formation of quality characteristics. Results indicated that the regulation of airflow significantly influenced water loss and the progression of withering, affecting internal physicochemical reactions in the leaves. Increasing airflow intensified water loss during withering. Comparative analysis under no-airflow conditions revealed a more even water distribution across the leaves. Airflow-assisted withering expedited water transfer and evaporation from stems to leaves or buds, converting other water phases into bound water, as revealed by LF-NMR and MRI. Dehydration stress altered the microscopic structures of the leaves, increasing stomatal conductance, but reducing it in leaves subjected to moderate airflow velocity in later withering stages. Increased water stress damages cellular structures, causing nutrient and enzyme leakage, stimulating enzyme activity and oxidative hydrolysis reactions that drive metabolite biosynthesis in finished tea. The no-airflow treatment favored the formation of key aroma difference compounds, such as (E)-2-hexen-1-ol and 1-hexanol, enriching catechins and caffeine and imparting a green and astringent character to the finished tea. Airflow-assisted withering enriched similar key aroma compounds, including linalool and its oxides, contributing to a floral foundation. The highest airflow treatment additionally enriched compounds such as phenylethyl alcohol, (E)-5,6-epoxy-β-ionone, and 3-hexen-1-ol, imparting a grassy aroma. This treatment also resulted in the enrichment of sugar compounds. The moderate airflow velocity treatments not only generated key difference compounds such as nerol but also significantly increased organic acid and amino acid contents, providing the finished tea with a fresh, floral aroma and a mellow, thick, and refreshing flavor. Moderate airflow velocity also promoted white tea color transformation to grayish-green color and significantly increased the immobilized water content. These findings suggested that the quality of finished tea withering at a moderate airflow velocity of 0.5 m/s surpassed that of the no airflow condition and the 1 m/s treatment. Moderate airflow velocity effectively controlled the withering process, enabling the gradual progression of oxidation, hydrolysis, and synthesis reactions and resulting in unique quality characteristics of white tea. These findings provide crucial evidence for the practical production processing of white tea and the operational parameters of withering equipment.

CRediT authorship contribution statement

Jinyuan Wang: Writing – original draft, Investigation, Formal analysis, Data curation, Conceptualization. Dongchun Lin: Investigation, Formal analysis, Data curation. Jiao Feng: Methodology, Investigation, Data curation. Shuping Ye: Investigation, Formal analysis. Weisu Tian: Software, Investigation. Weiyi Kong: Investigation, Formal analysis. Wenfeng Zhang: Software, Investigation. Qiang Chu: Visualization, Supervision, Conceptualization. Bugui Yu: Resources. Hongzheng Lin: Writing – review & editing, Conceptualization. Zhilong Hao: Writing – review & editing, Supervision, Project administration, Conceptualization.

Ethical statement

The authors ensure that the work described has been conducted in compliance with The Code of Ethics of the World Medical Association (Declaration of Helsinki) for experiments involving humans.

The ethical approval of sensory evaluation is not required by national laws. No human ethics committee or formal documentation process is available for sensory evaluation.

The authors affirm that appropriate protocols have been used to protect the rights and privacy of all participants. This includes ensuring that participation is voluntary, providing full disclosure of study requirements and risks, obtaining verbal consent from participants, not disclosing participant data without their knowledge, and allowing participants to withdraw from the study at any time.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

Supplementary material.

Data availability

Data will be made available on request.