Chemical analysis of ultrasonic-assisted treatment on improving cold-brewed green tea extraction efficiency and quality formation Authors: Chen Yang, Minghui Xu, Wei Jiang, Jiaqi Wang, Xu Zhong, Kaikai Wei, Guijie Chen, Qianying Dai, Chuanjian Cui, Ruyan Hou Published: 2026 ABSTRACT AbstractThis study investigated the feasibility of ultrasound-assisted cold brewing (UACB) for enhancing the efficiency of cold-brewed tea production while preserving its sensory qualities (color, aroma, taste). Scanning electron microscopy (SEM) revealed that UACB promoted tea leaf hydration and disrupted the surface morphology, providing more diffusion channels for metabolites. Metabolomics and quantitative analysis confirmed that UACB achieved higher extraction efficiency than cold-brewed, with alcohols, aldehydes, flavonoids, and chlorophylls exhibiting maximum increases of 3.91, 9.89, 1.53, and 1.43 fold, respectively. After 8 min of UACB, the extraction of major taste compounds such as L-theanine (55.46 μg/mL) and EGCG (72.10 μg/mL) was comparable to that after 1 h cold brewing (56.29 and 64.90 μg/mL, respectively). The rapid leaching of these compounds drives swift formation of floral and bean-like aroma, umami-sweetness, and green color of cold-brewed tea. These findings highlight UACB as an efficient method for enhancing cold-brewed tea production efficiency and quality control.Keywords: Cold-brewed tea, Ultrasound-assisted cold brewing, Metabolomics, Aroma, Taste, Color Introduction Tea is a flavored food product processed from the leaves of the Camellia sinensis plant (Zhai, Zhang, Granvogl, Ho, & Wan, 2022). For thousands of years, people have been dedicated to developing various processing and brewing methods to derive both flavorful enjoyment and health benefits from this remarkable leaf (Ye et al., 2022). Initially consumed as a medicinal decoction, tea gained worldwide popularity only after discovery that its processed leaves could produce a sensorially pleasing infusion when steeped in hot water (Zhao et al., 2025). Extensive studies confirm that tea's sensory attributes (color, aroma, and taste) originate from its diverse metabolite profile (e.g., alcohols, flavonoids, amino acids) in tea infusion (Zhai et al., 2022; Zhang, Cao, Granato, Xu, & Ho, 2020). As brewing parameters govern metabolite extraction efficiency, they directly determine ultimate tea quality (Liu et al., 2018). Consequently, elucidating flavor formation mechanisms and optimizing cold brewing protocols will facilitate tea product industrialization, advancing both tea chemistry research and commercial product excellence. Cold-brewed tea, as an emerging brewing technique, has seen its market exceed $2 billion in recent years, with green tea showing significant growth (Muller et al., 2020; Yang et al., 2025). Compared to conventional high-temperature extraction (>80 °C), low-temperature cold brewing (4 °C) enhances color vibrancy and floral aroma while reducing bitterness, driving its popularity among younger consumers (Yang et al., 2025). However, slow extraction kinetics at low temperatures hinder efficiency and scalability (Nancy et al., 2021; Yang, Wang, et al., 2025). Consequently, improving low-temperature extraction efficiency while preserving characteristic flavor metabolites has become a critical industry focus (Muller et al., 2020; Yang, Wang, et al., 2025). Ultrasound-assisted extraction (UAE) is a method that utilizes the cavitation effect generated by ultrasonic waves to rapidly disrupt plant cells, facilitating the quick release, diffusion, and dissolution of intracellular contents, thereby significantly improving extraction efficiency (Zhou et al., 2025). Traditional tea brewing relies on the concentration gradient between the rehydrated tea leaves and water to drive the release of tea metabolites. In contrast, during UAE, the jet impact formed by the asymmetric collapse of cavitation bubbles provides additional driving force, accelerating the extraction process (Gouda et al., 2021). Consequently, some studies have employed UAE at high temperatures to enhance the extraction efficiency of tea metabolites (Both, Chemat, & Strube, 2014). However, the feasibility of UAE under low-temperature conditions remains underexplored, and its impact on the extraction patterns of flavor compounds is still unclear. This study aims to explore the feasibility of ultrasound-assisted cold brewing (UACB) for enhancing the efficiency of cold-brewed tea production while preserving its sensory qualities. By integrating color and flavor profiling with metabolomics analysis, the dynamic changes in chemical composition and sensory attributes during different extraction stages were systematically examined. Furthermore, the microscopic mechanisms underlying UACB were elucidated from a microstructural perspective. These findings revealed the structural and chemical basis of UAE-enhanced cold brewing, providing theoretical support and methodological insights for improving the efficiency of cold brewing techniques. Materials and methods Tea samples and chemical reagent Anjibaicha green tea samples (AJBC) were procured in 2024 from Zhejiang Summit Angel Tea Co., Ltd. The key taste standard compounds used for quantitative analysis were purchased from Sigma-Aldrich Co., Ltd. (St. Louis, MO, USA). The detailed information of these standards was in the supplement. The n-alkane mixtures (C7-C30 straight-chain alkanes) were from Sigma-Aldrich, Inc.. The detailed information of formic acid, methanol and acetonitrile (mass spectrometry grade, LC-MS) was in the supplement. (−)-Epigallocatechin gallate (EGCG), (−)-epi-catechin gallate (ECG), (−)-epigallocatechin (EGC), (−)-epicatechin (EC), caffeine, and L-theanine were purchased from Sigma-Aldrich Co., Ltd. (St. Louis, MO, USA). NaCl (sodium chloride, analytically pure reagent) was purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Methanol and formic acid were purchased from Aladdin (Shanghai, China). The experimental water used was Wahaha purified water, purchased from Hangzhou Wahaha Co., Ltd. Tea infusions For all cold-brewed tea: the tea infusion made with a tea to-water ratio of 1:50 w/v, at 4 °C, for 4 min (C4M), 8 min (C6M), 16 min (C16M), 32 min (C32M), 44 min (C44M), 1 h (C1H), and 2 h (C2H) respectively. All UACB tea infusions were prepared with a tea-to-water ratio of 1:50 (w/v) and maintained at 4 °C during ultrasonic treatment. The treatment durations were 4 min (U4M), 8 min (U8M), 16 min (U16M), 32 min (U32M), 44 min (U44M),1 h (U1H), and 2 h (U2H), respectively. UACB experiments were performed using a FANGXU XI-600RT ultrasonic device (operating frequency: 40 kHz; amplitude: 1.2 μm; FANGXU, China) operated at a selected power of 60 W (Duan et al., 2026). An external cooling circulator (SENCO, China), preset at 4 °C, was connected to the ultrasonic system to continuously circulate chilled water through the ultrasonic bath, while the water circulation channel of the ultrasonic device remained open to facilitate heat dissipation. The ultrasonic bath was also maintained at 4 ± 0.5 °C, and the temperature of the ultrasonic medium was continuously monitored using a thermometer during sonication. All the green tea infusion samples were prepared in triplicate. Sensory quality assessment and quantitative descriptive analysis (QDA) For consistent sensory evaluation, all samples were assessed following the environmental requirements outlined in GB/T 23776-2018 (“Methodology for sensory evaluation of tea”). Environmental conditions for evaluation were set at (25 ± 1) °C and 40%–60% relative humidity. Neutral lighting and background colors were utilized. Furthermore, the space was adequately ventilated to prevent disturbances from ambient noise and odors. The sensory panel included 10 assessors (6 females, 4 males, aged between 22 and 54 years). Each participant had undergone over 2 years of professional sensory training and boasted more than 4 years of tea tasting experience. The panel members separately assessed the color, aroma, and taste of the green tea infusion samples with different treatments. For each sensory attribute of aroma and taste, they rated on a scale from 0 to 5 with increments of 0.1 (Yu et al., 2023a). Specifically, 0–1 indicated “barely detectable,” 1–2 meant “weak but recognizable,” 2–3 meant “perceptible and recognizable,” 3–4 meant “distinctly perceptible and clearly recognizable,” and 4–5 meant “strongly perceptible and strongly recognizable” (Sinija & Mishra, 2011; Yang, Wang, et al., 2025). All participants provided informed consent before the sensory evaluation after being fully briefed on the study objectives, procedures, potential risks, and their rights, including data confidentiality. Ethical approval for the sensory evaluation was obtained from the Ethics Committee of the College of Tea, Anhui Agricultural University, China. Data protection protocols were strictly implemented throughout the study. All materials complied with food safety standards and posed no risk to humans, animals, or the environment. Participants authorized the use of their sensory data, and the authors declare no conflict of interest. HS-SPME/GC–MS detection of volatile substances in cold-brewed and UACB green tea infusions The method was modified in accordance with previous studies (Yang, Wei, et al., 2025; Yun et al., 2021). 20 mL headspace vials (Agilent, Palo Alto, CA, USA) were employed to contain the green tea infusions. To enhance the extraction efficiency of volatile compounds from the tea infusion, 10 mL of tea infusion was added with 3 g of NaCl. The tea infusion was spiked with 10 μL of ethyl decanoate (1 ppm) and then shaken. To ensure the adequate volatilization of volatile compounds, a stir bar (1000 rotations per minute) was added to the 20-mL headspace vial. Headspace vials were placed in a 50 °C water bath for 10 min. Subsequently, a DVB/CAR/PDMS SPME fiber (50/30 μm, 1 cm; Supelco, Bellefonte, PA, USA) was inserted into the vial headspace and exposed for 30 min to adsorb volatile compounds. To desorb the volatile compounds adsorbed on the fiber, the SPME fiber was placed in the gas chromatography inlet (250 °C) for 5 min. The identification and semi-quantification of volatile organic chemicals (VOCs) were performed using an Agilent Model 8890 GC coupled with a 5977B mass spectrometer. The system was equipped with either a HP-5MS capillary column (30 m × 0.25 mm, 0.25 μm, J & W, Folsom, CA, USA) or a DB-WAX capillary column (30 m × 0.25 mm, 0.25 μm, J & W, Folsom, CA, USA). During the analysis, the DB-5MS column system was initially maintained at 40 °C for 5 min, after which the temperature was increased to 200 °C at a ramping rate of 4 °C/min and held for 1 min. It was then further elevated to 250 °C at 10 °C/min and kept at this level for 5 min. For the DB-WAX column system, the temperature program was as follows: 40 °C (maintained for 5 min), a rise from 40 to 200 °C at 4 °C/min (holding at 200 °C for 1 min), followed by an increase from 200 to 230 °C at 10 °C/min, with a final hold of 5 min at 230 °C. Helium (99.999% purity) was used as the carrier gas at a constant flow rate of 1 mL/min. The mass spectrometer operated under electron impact ionization at 70 eV, with mass spectra collected over a scan range of m/z 30–350. The ion source and quadrupole temperatures were set at 230 °C and 150 °C, respectively. The detailed methodology is provided in the Supplementary Materials. Retention indices (RI) for all detected volatiles were determined through linear interpolation using n-alkane retention times (C7-C30) as reference points. The detected compounds were compared with the information of compounds in the NIST database. Compound verification was supplemented by mass spectral fragments and m/z. Semi-quantitative analysis employed peak area ratios relative to ethyl decanoate (Yang, Wang, et al., 2025; Yang, Wei, et al., 2025). LC-QTOF-MS detection of non-volatile substances in cold-brewed and UACB tea green tea infusions Based on our previous method, an Agilent 1290 liquid chromatography-quadrupole time-of-flight mass spectrometry (LC-QTOF-MS) system coupled with a time-of-flight mass spectrometer was applied to analyze the difference of cold-brewed and UACB tea infusion samples (Yang et al., 2022; Yang, Wang, et al., 2025). An ACQUITY UPLC@HSS T3 Column (2.1 × 100 mm, 1.8 μm; Waters, Milford, MA, USA) was used to separate the compounds. The mobile phase A consisted of 0.1% formic acid in water, while the mobile phase B was acetonitrile. The flow rate was maintained at 0.3 mL/min, with the column temperature set at 35 °C. The sample tray was kept at room temperature, and the injection volume was 3.0 μL. The gradient elution program was as follows: 0 min, 5% B; 3 min, 10% B; 12 min, 20% B; 14 min, 25% B; 16 min, 95% B; 18 min, 95% B; 20 min, 5% B; 22 min, 5% B. The mass spectrometry parameters were set as follows: sheath gas flow rate at 11 L/min, auxiliary gas flow rate at 8 L/min, nebulizer pressure at 35 psi, gas temperature at 320 °C, sheath gas temperature at 350 °C, and capillary voltage at 3500 V. In negative ion mode, the mass scan range was set from m/z 100 to 1300. UPLC-QQQ-MS/MS detection of major taste substances in cold-brewed and ultrasound-assisted cold extraction tea green tea infusions According to previous studies, key taste compounds were quantitatively analyzed using an ExionLC system coupled to a triple quadrupole linear ion trap mass spectrometer with electrospray ionization (ESI) (Yang, Wang, et al., 2025). The target compounds in tea infusion samples were quantified through an external calibration method. The chromatographic separation was performed on an Agilent C18 column (2.1 × 50 mm, 1.8 μm) at a flow rate of 0.3 mL/min and a temperature of 40 °C. The injection volume was set at 3 μL, and the elution gradients of mobile phases A (0.1% formic acid-water, v/v) and B (acetonitrile) were as follows: 0–1 min, 99.5% A; 1–5 min, 99.5%–95% A; 5–10 min, 95% A; 10–12 min, 95%–93% A; 12–16 min, 93%–90% A; 16–31 min, 90%–75% A; 31–32 min, 75% A; 32–33 min, 75%–99.5% A; 33–35 min, 99.5% A. The elution gradients were as follows. The mass spectrometry conditions were set as follows: ion spray voltage at −4500 V, turbo spray temperature at 500 °C, nebulizer gas at 55 psi, heater gas at 55 psi, and curtain gas at 30 psi. In the selected ion monitoring mode, the production pair for each standard was selected, and its collision energy (CE) and declustering potential (DP) were optimized. All samples and standards were analyzed in multiple reaction monitoring (MRM) mode. Chromaticity measurement The chromatic characteristics (L⁎, a⁎, and b⁎ values) of the tea infusions were measured using a Konica Minolta benchtop spectrophotometer, with samples placed in a transparent rectangular chamber (50 × 38 mm, 10 mm optical path length). The L⁎ value reflected changes in brightness, extending from black to white (+white, −black). The a⁎ value showed the difference between red and green tones (+red, −green). The b⁎ value illustrated the contrast between yellow and blue shades (+yellow, −blue). Kinetic analysis of major taste compounds The BoxLucas1 model was employed to fit the first-order kinetic model of major taste compounds (L-theanine, caffeine, gallic acid, EGCG, EGC, ECG, EC, C) in UACB and cold-brewed green tea infusions at extraction times of 4, 8, 60, and 120 min. Qt represents the concentration of the taste compounds at time t (μg/mL); t represents the extraction time (min); k represents the reaction rate constant; and Qe represents the final concentration of the taste compounds (μg/mL). Scanning electron microscopy (SEM) SEM (Quanta FEG 250, Hefei, China) observed the morphology of eight different kinds of treated green tea leaves. Green tea leaves samples were fixed onto SEM stubs, and then observed at an accelerating voltage of 20 kv. The samples were sprayed with platinum to increase their electrical conductivity on a LEICA EM SCD 500 at 16 mA with 120 s. All green tea samples were imaged at a uniform magnification of 500×, with a scale bar of 20 μm for all images. Statistical analysis GraphPad Prism (version 10.0) was used to perform data processing and bar chart generation. MS-Convert converts LC-MS raw data format to mzML, and MS-DIAL (version 5) is used on converted data to filter noise, identify peaks, analyze overlapping peaks, align peaks, and fill peaks (detailed parameters were presented in Table S1) (Yang et al., 2022). Principal component analysis (PCA) and orthogonal partial least squares discriminant analysis (OPLS-DA) were conducted using SIMCA 14.1 software. The heatmaps were constructed using TBtools software. The significance analysis was performed using SPSS statistical software (version 14.0; SPSS Inc., Chicago, IL). Results and discussions Sensory evaluation of key stages in flavor formation of cold-brewed and UACB green tea To determine the critical time points for flavor formation in cold-brewed and UACB green tea, sensory evaluations were conducted on tea infusion samples extracted for 4 min, 8 min, 16 min, 32 min, 44 min, 1 h, and 2 h (Fig. S1). Under cold-brewing conditions, the aroma and taste intensity of the tea infusion gradually increased with prolonged extraction time. After 1 h of extraction, the tea infusion exhibited the unique sensory characteristics typical of cold-brewed green tea, indicating that the substantial generation of flavor compounds primarily occurred after this time point (Yang, Wang, et al., 2025). The sensory evaluation results of the UACB tea infusion showed that the intensities of most aroma characteristics, such as green, malty, floral, cooked-vegetable, and bean-like aromas, in the tea infusion extracted for 8 min were comparable to those of the cold-brewed tea infusion extracted for 1 h. Compared to the 4 min extraction, the UACB process significantly enhanced key aroma and taste attributes: green, malty, bean-like, umami, and sweetness (p < 0.05). However, some aroma characteristics (e.g., green aroma) gradually weakened with the progression of extraction after 16 min. Based on the sensory evaluation results, 4 min and 8 min were identified as critical time points for flavor formation in UACB, while 1 h and 2 h were identified as critical time points for flavor formation in cold brewing. Therefore, further quality analyses were primarily conducted on tea infusion samples extracted for 4 min, 8 min, 1 h, and 2 h. Effect of UACB on the leaf morphology and tea infusion color Effect of UACB on the leaf morphology The morphological evolution of tea leaves and the tea infusion during UACB was initially captured and characterized using digital imaging techniques, with representative visual documentation provided in Fig. 1. As shown in Fig. 1a, the tea leaves gradually unfurled with increasing extraction time. Under identical processing duration, the UACB method induced significantly greater leaf expansion compared to conventional cold-brewed. This difference was particularly pronounced during shorter steeping periods (less than 8 min). The leaching of chemical composition from tea leaves generally involves water absorption and expansion, dissolution of cellular constituents, and subsequent diffusion driven by concentration gradients (Yang, Wei, et al., 2025). The extent of leaf unfurling during hydration can serve as a macroscopic indicator of enhanced leaching efficiency, as it reflects the increased mass transfer rates within the leaf structure. Dynamic changes in tea leaf appearance, infusion color, and sensory attributes during conventional cold brewing and ultrasound-assisted cold brewing (UACB). (a) Appearance of infused tea leaves at different extraction times: C4M, C8M, C1H, C2H (cold brewing for 4 min, 8 min, 1 h, 2 h) and U4M, U8M, U1H, U2H (UACB for 4 min, 8 min, 1 h, 2 h). (b) Appearance of tea infusions corresponding to the same time points. (c) Aroma profiles of cold-brewed tea infusions (radar plot). (d) Aroma profiles of UACB tea infusions. (e) Taste profiles of cold-brewed tea infusions. (f) Taste profiles of UACB tea infusions. (g–i) Colorimetric parameters: (g) L⁎ (lightness), (h) a⁎ (redness/greenness), (i) b⁎ (yellowness/blueness) for both brewing methods. Error bars with different letters indicate significant differences according to one-way ANOVA with Duncan and LSD test (p < 0.05). Effect of UACB on the color of tea infusion The observed color changes in the tea infusion further supported this observation. As shown in Fig. 1b, the tea infusion gradually darkened with prolonged extraction time, with the UACB treatment exhibiting faster color development and visually more intense coloration compared to conventional cold-brewed. To quantify these color variations, we analyzed the changes in color parameters L⁎ (lightness), a⁎ (redness), and b⁎ (yellowness) during the brewing process. The results showed that as brewing time increased, the cold-brewed tea infusion exhibited a decrease in L⁎ from 97.12 (4 min) to 94.96 (2 h), a⁎ from −3.22 (4 min) to −4.70 (2 h), and an increase in b⁎ from 10.58 (4 min) to 19.80 (2 h). In contrast, the UACB tea infusion showed a more pronounced change: L⁎ decreased from 95.64 (4 min) to 89.04 (2 h), a⁎ decreased from −3.88 (4 min) to −6.32 (1 h), and b⁎ increased from 16.06 (4 min) to 37.07 (2 h). These results indicate that prolonged brewing time leads to progressive darkening of the tea infusion (decreased L⁎), accompanied by a shift toward a more yellowish-green hue (increased b⁎ and decreased a⁎). Notably, the UACB tea infusion exhibited darker coloration and higher saturation than the cold-brewed tea, demonstrating that ultrasound assistance may enhance extraction efficiency at low temperatures, and produce a more concentrated tea infusion. Effect of UACB on the microstructure of tea leaves To elucidate how UACB-induced morphological changes affect tea quality, SEM was employed to examine the microstructural evolution of tea leaves during brewing. Fig. 2a–d shows tea leaves subjected to conventional cold brewing at different extraction times. After 4 min (Fig. 2a), the leaf surface remained intact with no visible signs of expansion. At 8 min (Fig. 2b), the surface became slightly smoother, but the overall structure was largely preserved. After 1 h (Fig. 2c), the leaf surface turned noticeably smooth due to water absorption. By 2 h (Fig. 2d), the leaching of soluble constituents exposed stomatal structures. This slow hydration is attributed to the reduced thermal motion of water molecules at low temperatures (Zhang et al., 2020) and the rigidity of cell wall components (cellulose and hemicellulose) (Luo, Huang, Chang, Li, & Guo, 2023), resulting in a slower rate of water absorption and cell expansion compared to high-temperature environments. Consequently, extraction efficiency was reduced, and visible color development in the tea infusion became apparent only after 1 h of cold brewing (Fig. 1b). Scanning electron microscopy (SEM) images of tea leaf surfaces after different brewing treatments (magnification 500×; scale bar = 20 μm for all images). (a–d) Conventional cold brewing: (a) C4M (4 min), (b) C8M (8 min), (c) C1H (1 h), (d) C2H (2 h). (e–h) Ultrasound-assisted cold brewing (UACB): (e) U4M (4 min), (f) U8M (8 min), (g) U1H (1 h), (h) U2H (2 h). Arrows in (e–h) highlight surface disruption and increased porosity due to cavitation. In contrast, UACB induced rapid microstructural changes. Fig. 2e–h show tea leaves subjected to UACB. After only 4 min (Fig. 2e), the leaf surface already exhibited cracks and localized damage. At 8 min (Fig. 2f), more extensive surface disruption and increased porosity were observed. After 1 h (Fig. 2g), the leaf surface became highly fragmented, and by 2 h (Fig. 2h), the original cuticular structure was largely destroyed. Previous studies on UAE in other plant matrices have suggested that the microjets generated by ultrasound can flush the surface of materials, reduce the boundary layer resistance, and promote faster water penetration into the interior tissues (Hallouch et al., 2025; Shen et al., 2023), which may explain the accelerated expansion rate observed in this study. Furthermore, cavitation shock waves caused substantial structural disruption on the tea leaf surface (Fig. 2e-h), greatly increasing the contact area between active components and water. Beyond these mechanical effects, the transient collapse of cavitation bubbles generates localized high temperatures. Although the brewing solution remains at 4 °C, these microsecond-scale high-temperature events reduce the viscosity of the boundary layer and enhance molecular desorption from the leaf matrix, thereby increasing the effective diffusion coefficient and the overall mass transfer driving force (Shen et al., 2023). As a result, UACB alleviates the restrictive effect of cell wall rigidity at low temperatures and accelerates the release of intracellular substances. Consistent with these observations, the UACB-treated tea infusion (8 min) showed L⁎, a⁎, b⁎ values similar to or higher than those of the 2-h cold-brewed infusion (Fig. 1b). Effect of ultrasound-assisted cold brewing on volatile components in green tea infusion Multivariate statistical analysis and the changes in volatile components Aroma characteristics critically determine tea infusion quality (Kun et al., 2025). This study thus applied HS-SPME/GC–MS and sensory analysis to systematically compare the aromatic profiles and volatile dynamics of tea infusions prepared by UACB and traditional cold-brewed. According to the detection results of HS-SPME/GC–MS, a total of 78 volatile compounds were detected under the two brewing conditions of cold-brewed and UACB for green tea. Table S2 shows the detailed detection results, including 22 alcohol compounds, 16 aldehyde compounds, 12 ketone compounds, 14 ester compounds, 3 terpenes, 3 epoxides, 4 nitrogen-containing heterocyclic compounds, and 4 other types of compounds. Among them, alcohols, aldehydes, and esters were the main volatile compounds in both cold-brewed and UACB green teas (Fig. 3a). Identification and multivariate analysis of volatile compounds in cold-brewed and UACB green tea infusions. (a) Circular diagram showing the distribution of volatile compound classes. (b) Orthogonal partial least squares discriminant analysis (OPLS-DA) score plot discriminating samples by brewing method and time. (c) Hierarchical cluster analysis (HCA) dendrogram showing grouping of samples based on volatile profiles. (d) Venn diagram illustrating the number of unique and shared volatile compounds between cold-brewed (C) and UACB (U) samples. Abbreviations: C4M, C8M, C1H, and C2H denote cold brewing for 4 min, 8 min, 1 h, and 2 h, respectively; U4M, U8M, U1H, and U2H denote UACB for 4 min, 8 min, 1 h, and 2 h, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) The characteristic role of UACB in shaping green tea infusion aroma at low temperatures was further investigated by performing OPLS-DA and HCA on GC–MS detection results. The reliability of the OPLS-DA model was confirmed, with satisfactory model fitting (R2Y = 0.988), and predictive ability (Q2 = 0.944). The results of the OPLS-DA clustering analysis are shown in Fig. 3b. In the cold-brewed samples, C2H sample was separated from the samples brewed for other extraction times. In the UACB samples, those extracted for 8 min and 1 h were differentiated from the others. HCA further revealed that UACB-treated samples at 8 min showed aroma profiles approaching those of 2 h cold-brewed tea (Fig. 1c). Specifically, the intensities of green, floral, and bean-like notes were comparable between them (Fig. 1d). This clustering result supports the OPLS-DA finding and highlights the time-saving advantage of UACB (Wen et al., 2025). In summary, the distinct separation observed in the OPLS-DA analysis and the consistent clustering results obtained from the HCA analysis collectively demonstrate the reliability and accuracy of HS-SPME/GC–MS analysis as a method to distinguish different cold-extracted green tea samples. The Venn diagram (Fig. 3d) further elucidated the relationships among these samples. The results indicated that, among the eight tea infusions subjected to cold-brewed and UACB, the majority of compounds were overlapping. Further comparison of these compounds revealed that only butyl acrylate and α-terpinene were only present exclusively in the tea infusions obtained by UACB and cold-brewed, respectively. The majority of the compounds were common to both extraction methods. Therefore, it could be inferred that the UACB method, which maintained a low temperature, retained a similar tendency in types of compound extraction as cold brewing (Yang, Wang, et al., 2025). Dynamic changes of aroma quality in UACB tea infusion Despite the fact that neither cold brewing nor UACB led to the formation of new compounds in green tea infusion, the tea infusions produced by these two methods still exhibited different aroma profiles (Fig. 1c, d). For instance, compared with cold-brewed tea, the ultrasonic-assisted extraction method (U8M, U1H) had more pronounced and intense green, malty, bean-like, and mothball aromas. According the results of venn diagram findings, it was further speculated that the differences in aroma profiles were due to the differences in the accumulation concentrations of the compounds (Fig. 4a). Proportions and heatmaps of volatile compounds in cold-brewed and UACB tea infusions. (a) Stacked bar chart showing the relative proportions of different volatile compound classes across extraction times for both methods. (b–f) Heatmaps of individual compounds: (b) alcohols, (c) ketones, (d) aldehydes, (e) esters, (f) terpenes and other compounds. Color intensity reflects normalized concentration (low = blue, high = red). Abbreviations: C4M, C8M, C1H, and C2H denote cold brewing for 4 min, 8 min, 1 h, and 2 h, respectively; U4M, U8M, U1H, and U2H denote UACB for 4 min, 8 min, 1 h, and 2 h, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) According to the Venn diagram analysis (Fig. 4a), alcohols and aldehydes were the two largest categories of compounds in the samples, and the concentrations of these compounds varied under different extraction conditions. For these two compound classes, UACB achieved higher extraction efficiency than cold brewing, with significant concentration differences at corresponding time points (p < 0.05). Under low temperature conditions, the concentration of compounds gradually increased with the extension of extraction time. In contrast, during the UACB processes, the extraction of aromatic compounds first increased and then decreased with the extension of extraction time. This is probably caused by the accelerated volatilization of these compounds induced by ultrasonic cavitation and microstreaming (Duan et al., 2026). Specifically, the concentration of aromatic compounds extracted by ultrasonic treatment for 1 h was the highest. A reduction in total volatile compound concentration was observed following 1 h of processing. This might be the key point for the decrease in green and cooked-vegetable notes after 8 min of UACB, as well as the reduction in fruity, malty, floral, and mothball notes after 1 h (Fig. 1c, d). Alcohols Alcohols are typically associated with floral and green aroma attributes in green tea (Yin et al., 2022; Zhai et al., 2022). Prolonged low-temperature extraction led to the accumulation of higher concentrations of linalool and geraniol, which were identified as key factors in enhancing the floral aroma of cold-brewed tea (Yu et al., 2023b). Similarly, UACB achieved a significant accumulation of alcohol compounds within a short extraction time. As shown in Fig. 4b, the total alcohol content in UACB tea infusion reached 13.51 ± 0.88 μg/L at 4 min and 54.23 ± 1.72 μg/L at 8 min, which were significantly higher than those in conventional cold-brewed tea at the corresponding time points (C4M: 6.51 ± 0.91 μg/L; U8M: 13.88 ± 0.49 μg/L) (p < 0.05; Fig. 4b). After 8 min of extraction, the concentrations of individual alcohols in the UACB samples peaked, including 1-pentene-3-ol (green), 1-pentanol (fruit), (Z)-3-hexen-1-ol (green fruit), (E)-2-hexen-1-ol (green fruit), 1-hexanol (flower), 1-heptanol, 1-octen-3-ol (mushroom), 2-ethyl-1-hexanol (green), benzyl alcohol (rose), 1-octanol (almond), linalool (floral), phenylethyl alcohol (floral), 1-nonanol (floral), (Z)-geraniol (floral), geraniol (lemon, floral) (Yang, Wei, et al., 2025; Zhai et al., 2022). Among these, benzyl alcohol, linalool, 1-nonanol, (Z)-geraniol, geraniol and phenylethyl alcohol have been confirmed as key alcohol compounds influencing the floral and overall aroma profiles of most green teas (Xie et al., 2023; Yu et al., 2023a). Additionally, (Z)-3-hexen-1-ol, linalool, and geraniol are proven key factors contributing to the floral aroma quality of Taiping Houkui green tea (Jin et al., 2024). Consequently, the accumulation of these compounds explains the enhanced floral, fruity, and green aromas observed in the UACB (8 min) tea infusion compared with its cold-brewed infusion (p < 0.05; Fig. 1c, d). This rapid extraction effect is likely attributed to the localized high-temperature and high-pressure conditions generated by the collapse of cavitation bubbles during ultrasonic extraction, which promote the release of alcohols. However, after 1 h of extraction, a significant decrease in concentration was observed. This decline may be due to the fact that prolonged ultrasonic treatment enhances the molecular motion of highly volatile alcohols, increasing their volatility and leading to a marked loss, even under low-temperature conditions. Aldehydes Aldehyde compounds in green tea are derived from the degradation and transformation of precursor substances such as fatty acids, amino acids, carotenoids, and glycosides during green tea processing (Zhai et al., 2022). Typically, these aldehyde compounds are considered key contributors to the malt-like, floral, and grassy aromas of green tea. Previous studies have found that aldehydes in green tea are rapidly extracted during hot water brewing, with their cumulative concentrations being higher than those in cold-brewed green tea (Yang, Wang, et al., 2025). This study reveals that under UACB conditions, aldehyde compounds in green tea are rapidly extracted within a short period (U8M: 22.02 ± 0.43 μg/L). This suggested that the ultrasonic field facilitated the dissolution of aldehydes and enhanced their diffusion behavior. Furthermore, previous studies had identified several aldehyde compounds as key odor-active constituents responsible for the bean-like and chestnut-like aromas in green tea. These compounds included 3-methylbutanal, heptanal, benzaldehyde, octanal, benzeneacetaldehyde, decanal, nonanal, and (E)-2-octenal (Zhang et al., 2024; Zhu et al., 2018). The relatively higher accumulation of these compounds under UACB conditions might account for the more pronounced bean-like aroma in the tea infusion (Fig. 1d). Ultrasonic extraction technology helped mitigate the limitation of insufficient extraction of bean-like aroma compounds from green tea during low-temperature processes, thereby enhancing the typical aroma attributes of green tea. However, similar to alcohols, most aldehydes also exhibit a process of increasing and then decreasing overall concentrations under UACB conditions. This trend may be attributed to both volatilization and slow oxidation reactions, which can occur even at low temperatures, as exposure to oxygen coupled with ultrasound-induced enhancement of intermolecular interactions promotes these processes (Yu et al., 2025). This may explain why the tea infusion under UACB conditions has more pronounced malty and green aroma attributes in the first 1 h of extraction compared with cold brewing, but a decrease in malty aroma concentration occurs after 2 h. Ketones Ketone compounds impart floral and fruity aroma attributes to green tea (Zhai et al., 2022). In this study, the detected ketone aroma compounds accounted for 7.98% of the total relative content. Except for 6,10-dimethyl-5,9-undecadien-2-one, all ketone compounds reached their highest concentrations in the samples extracted by UACB for 8 min and 1 h. These compounds contained (E)-β-ionone (floral), which has been identified as a key aroma compound contributing to the floral quality of green tea (Yu et al., 2023b; Zhai et al., 2022). Its OAVs in U8M and U1H were 24 and 23, respectively, further confirming its important contribution to the floral aroma profile of UACB green tea infusions. Camphor also reached its highest concentration in the 1 h UACB green tea sample, making this sample exhibit the most pronounced mothball-like aroma (Yu et al., 2023b). The dynamic changes of these compounds indicated that samples extracted for 8 min to 1 h by ultrasound had more pronounced floral and mothball-like odor attributes. However, it is noteworthy that prolonged ultrasonic treatment accelerated the volatilization of these compounds, leading to a decline in their aroma attributes (Figs. 1d, 4c). Esters Esters contribute intense floral and fruity aromas to tea infusions (Zhai et al., 2022). Previous studies indicated that ester compounds are more easily extracted during short-term hot brewing but need more time to accumulate under low-temperature conditions (Yang, Wei, et al., 2025). This study demonstrated that UACB accelerated the extraction of esters even at low temperatures. The total relative content of ester compounds in the U8M sample (14.38 μg/L) was significantly higher than that in the cold-brewed sample (3.21 μg/L, p < 0.05), indicating rapid ester accumulation within just 8 min. Furthermore, the concentrations of key ester compounds, such as (Z)-3-hexenyl butyrate (fruity note) and methyl salicylate (floral note), in the U8M tea infusion were higher than those in the 2 h cold-brewed tea (p < 0.05; Fig. 4a). This enhancement further intensifies the floral and fruity aroma profile of cold-brewed green tea. Terpenes and others Three terpene compounds were detected in this study. Among them, β-myrcene (geranium-like) and (E)-β-ocymene (floral) exhibited higher solubility contents in the UAE method, similar to alcohol compounds. However, α-terpinene was found exclusively in the tea infusion obtained by cold extraction. Although ultrasound generally enhances the extraction of volatile compounds, the high volatility of α-terpinene may lead to its partial evaporation under the localized high-pressure and high-temperature conditions generated by ultrasound, limiting its retention even under low-temperature conditions. The epoxy compounds, which are mainly linalool oxides, including (Z)-linalool oxide (furanoid), (E)-linalool oxide (furanoid), and (E)-linalool oxide (pyranoid), have been proven to be key aroma compounds contributing to the floral quality of green tea. It was found that during the overall extraction process, these three epoxy compounds could be enriched and reached their peaks within 1 h in the ultrasonic-assisted cold extraction environment. This has a positive impact on enhancing the floral aroma quality of green tea. In this study, a total of four nitrogen-containing heterocyclic compounds were detected. Indole, characterized by a floral aroma, was the representative compound in green tea (Jin et al., 2024). Indole was detected in both the tea infusions obtained by UACB for 8 min and by cold brewing for 2 h. The results indicated that indole requires a longer extraction time under cold brewing conditions to be effectively extracted (Fig. 1c). According to the past studies (Yang, Wang, et al., 2025), indole was more readily extracted in a hot water compared. Unlike cold brewing, indole can be rapidly eluted within 8 min with the assistance of ultrasound. It is noteworthy that indole was not detected in subsequent time points, likely due to its rapid volatilization caused by the cavitation effect of ultrasound. The other three nitrogen-containing heterocyclic compounds, namely 1-ethyl-1H-pyrrole, 1-ethyl-1H-pyrrole-2-carboxaldehyde, and 2,5-dimethyl-3-ethylpyrazine, exhibited caramel, nutty, and roasted-like aromas, respectively (Wang et al., 2022). Past studies had shown that 2,5-dimethyl-3-ethylpyrazine was also identified as a key aroma compound contributing to the bean-like aroma in green tea (Wang et al., 2022). Compared with cold brewing, ultrasound demonstrated a greater efficacy to extract 1-ethyl-1 h-pyrrole-2-carboxaldehyde and 2,5-dimethyl-3-ethylpyrazine in a short time under low-temperature conditions. Their accumulation had a positive effect on enhancing the bean aroma of green tea infusion in the ultrasonic environment. Among other types of compounds, dimethyl sulfide, a key aroma-active compound in green tea, was detected (Jin et al., 2024). Dimethyl sulfide is a highly volatile substance generated from the thermal decomposition of S-methylmethionine (SMM) during tea processing. It has been confirmed in studies on the floral aroma quality of Lu'an Guapian green tea that dimethyl sulfide plays a crucial role in the formation of floral aroma in green tea (Yu et al., 2023a). Throughout the low-temperature ultrasonic process, an ultrasonic time of less than 1 h did not lead to the volatilization and degradation of dimethyl sulfide. Instead, it accumulated in the tea infusion to the highest concentration within this time frame, which is in contrast to the pattern observed in hot brewing. This contributed to the enhancement of the overall aroma quality of green tea. Effect of UACB on non-volatile compounds in tea infusion Multivariate statistical analysis and the changes in non-volatile components Non-volatile compounds are key contributors to the taste profile of tea infusions. Therefore, an LC-MS-based untargeted metabolomics approach was employed to analyze the compositional changes and quality-contributing characteristics of these compounds in tea infusion. After data preprocessing with LC-MS and MS-DIAL, a total of 3091 ion features were obtained for multivariate statistical analysis (PCA, OPLS-DA). The results of PCA showed that all green tea samples from cold extraction and ultrasonic-assisted cold extraction fell within the 95% confidence interval. The first and second principal components accounted for 80.10% and 12.10% of the total variance, respectively, with a cumulative contribution rate of 92.20%. This indicated that the dimensionality-reduced data could explain most of the characteristics of these green tea samples (Wen et al., 2025). The results presented in Fig. 5a indicate that all green tea samples can be differentiated based on PC1. The green tea samples subjected to ultrasonic-assisted cold extraction for less than 1 h and cold extraction for less than 2 h cluster together, suggesting that the differences among these samples account for only a small proportion of the total variance. In contrast, the 1 h and 2 h UACB samples clustered together on one side of PC2, while the 2 h cold-brewed sample was separated on the opposite side. To further analyze the variation between UACB and cold-brewed green tea, an OPLS-DA model was subsequently established. As illustrated in Fig. 5b, the model successfully differentiated cold-brewed green tea from UACB-processed samples, demonstrating robust fitting performance and reliable predictive accuracy (R2 = 0.95, Q2 = 0.85). The metabolites were marked based on their variable importance in projection (VIP) scores, with those exceeding a VIP threshold of 1.0 being identified as potential marker compounds (Table S3). A total of 41 key constituents were characterized through analysis of mass fragmentation patterns, comparison with reference standards, and matching against established tea metabolite databases. Compounds such as flavones (bitterness), flavone glycosides (bitterness), flavanols (bitterness), phenolic acids (astringency and astringent-bitter), organic acids (sourness), and alkaloids (bitterness) were found to increase with brewing time both in cold-brewed and UACB, with differences in their relative contents. The heatmap illustrated the variations in compound concentrations of these metabolites across different brew processes (Fig. 5c-f). Multivariate analysis and heatmaps of non-volatile compounds in cold-brewed and UACB green tea infusions. (a) Principal component analysis (PCA) score plot. (b) OPLS-DA score plot discriminating cold-brewed and UACB samples. (c) Heatmaps of selected non-volatile metabolites. Abbreviations: C4M, C8M, C1H, and C2H denote cold brewing for 4 min, 8 min, 1 h, and 2 h, respectively; U4M, U8M, U1H, and U2H denote UACB for 4 min, 8 min, 1 h, and 2 h, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) Dynamic changes of taste quality in UACB tea infusion Catechins and their derivates Catechins and their derivatives are widely recognized as the primary contributors to the bitterness and astringency of green tea. The dynamic changes in key catechin compounds, including (−)-epiafzelechin, (+)-catechin, EGC, (−)-gallocatechin, and the catechin dimer procyanidin dimer (EC-ECG)—during UACB and cold brewing are illustrated in Fig. 5c. Due to the relative stability of the C-3 ester bond in catechins under low-temperature conditions, the concentrations of these compounds increased with prolonged extraction time under both methods. Notably, the catechin content in UACB tea infusion after 8 min was higher than that in cold-brewed tea at 8 min but slightly lower than after 1 h of cold brewing (Yang, Wei, et al., 2025). However, when UACB exceeded 1 h, the catechin content surpassed that of cold brewing after 2 h. This indicates that UACB maintains relatively low bitterness/astringency in short durations (8 min), whereas prolonged extraction (1 h) leads to excessive catechin accumulation and intensified harshness, consistent with sensory evaluation results (Fig. 1e, f). Additionally, faster formation of catechin polymers (e.g., EC-ECG) was observed during UACB, indicating that ultrasonic energy may promote catechin polymerization and related transformations (Duan et al., 2026). These derivatives typically contribute yellowish to the tea infusion, explaining the darker color of UACB tea compared to traditional cold brewing (Fig. 1b) (Long et al., 2024). Nevertheless, the catechin transformation rate remained lower than in conventional hot brewing, implying that ultrasonic energy alone is insufficient to meet the thermodynamic free energy required for extensive catechin oxidation. These findings demonstrate that UACB not only accelerates the formation of flavor compounds but also stabilizes their composition in shorter durations. Flavonoids and flavonol glycosides Flavonoids and flavonol glycosides have been identified as key components responsible for the yellow coloration and astringency of green tea infusions (Ye et al., 2022). Fig. 5 illustrates the dynamic extraction profiles of six critical differential flavonoids (kaempferol and rutin) and their glycosides (kaempferol-O-robinobioside, kaempferol-O-galloylhexoside-1) under UACB and cold brewing conditions at different steeping durations. The results demonstrate that cold brewing yields slower leaching kinetics for both flavonoids and flavonoid glycosides, while UACB exhibits accelerated release dynamics. This phenomenon can be attributed to the mechanical disruption of tea leaf cellular structures by ultrasound, as evidenced by SEM observations (Fig. 2), which enhances the dissolution of flavonoids during sonication. Notably, kaempferol and kaempferol-O-robinobioside play pivotal roles in shaping the tea infusion's astringency and yellow color quality (Wen et al., 2025), which explains why prolonged UACB duration leads to markedly intensified astringency and increased b⁎ values in the tea infusions. Amino acids and soluble sugars For cold-brewed green tea, the most characteristic flavor profile was its umami-sweet taste, with amino acids and soluble sugars serving as the key compounds responsible for this quality (Long et al., 2024). In traditional hot brewing, amino acids readily react with catechins under high temperatures to form new compounds or decompose into aromatic substances. (Han et al., 2022), ultimately diminishing the umami perception of the tea infusion. This study findings indicate that the additional energy input from ultrasound may not induce chemical transformations of amino acids and soluble sugars. As illustrated in Fig. 5c, under both UACB and cold brewing conditions, three amino acids (γ-aminobutyric acid, L-theanine, and l-glutamine) and soluble sugars exhibited time-dependent accumulation, achieving rapid extraction within just 8 min. This explains why the tea infusion develops a relatively pronounced umami and sweet taste profile in such a short time. However, after 1 h, the extraction rates of amino acids and soluble sugars began to slow, while bitter-tasting compounds (e.g., catechins and flavonol glycosides) continued to accumulate. As a result, the relative contribution of umami and sweet flavors appeared to be gradually masked by these bitterness and astringent components. Phenolic acids and alkaloids Phenolic acids are among the key compounds influencing the astringency of tea infusion (Ye et al., 2022). This study identified phenolic acids (e.g., quinic acid and vanillic acid) and their derivatives (e.g., quinic acid derivative-3 and galloyl-HHDP-glucose) as the major differential metabolites between UACB and cold-brewed tea. The concentrations of these phenolic acids increased with prolonged extraction time, explaining the continued intensification of astringency in both brewing methods after extended steeping. Furthermore, phenolic acids in tea infusion undergo oxidation reactions, generating colored compounds such as quinones, which darken the tea infusion. Notably, phenolic acids with multiple hydroxyl groups are prone to oxidation by air or reactions with other components, forming phenolic acid derivatives that reduce tea infusion brightness and induce yellowing (Long et al., 2024). Ultrasound may accelerate the conversion of phenolic acids, resulting in a more pronounced yellowish hue in the tea infusion. Chlorins and related compounds Chlorins and their derivatives play a crucial role in determining the characteristic color properties of green tea infusions (Wen et al., 2025). Previous studies indicate that chlorophyll a and chlorophyll b impart bluish-green and yellowish-green hues to tea infusions, respectively. Chlorophyllin, a chlorophyll derivative, exhibits a distinct green color due to its structural characteristics. During the UACB process, the contents of chlorophyllide a and chlorophyllide b gradually increased. Additionally, higher levels of pheophorbide a and pheophorbide b were observed in the tea liquor during UACB, indicating that ultrasonic energy input accelerates the conversion of chlorophyll. The accumulation of these compounds contributed blue-green and yellow-green hues to the tea infusion (Fig. 1h–i), thereby progressively shaping the appearance of the tea infusion color in cold-brewed tea (Yu et al., 2019). Alkaloids are key compounds responsible for the bitter taste of tea infusions. Heatmap analysis showed that the concentrations of two alkaloids, camellimidazole A (bitterness) and camellimidazole B (bitterness), were relatively higher in ultrasound-assisted samples and gradually increased with prolonged ultrasonic treatment (Wen et al., 2023). This trend was consistent with the changes in bitterness intensity observed during the sensory evaluation (Fig. 1e, f). This may explain why UACB yields a more pronounced bitterness compared to conventional cold brewing. Quantitative and kinetic analysis of major taste compounds in green tea during UACB Integrated analysis of sensory evaluation data (Fig. 1e, f) revealed that both UACB and conventional cold brewing produced infusions characterized by low bitterness and pronounced umami sweetness. Notably, under ultrasound-assisted conditions, the characteristic taste profile of cold-brewed tea developed more rapidly. Therefore, this study conducted quantitative and first-order kinetic model analysis of major taste compounds in green tea infusions (Fig. 6, Fig. S2). Quantitative changes of major non-volatile taste compounds in cold-brewed and UACB green tea infusions over brewing time. Bar plots show mean concentrations (μg/mL) with error bars (SD). (a) L-theanine, (b) EGCG, (c) ECG, (d) caffeine, (e) EGC, (f) EC. * indicated significance at p < 0.05; ** indicated significant at p < 0.01; *** indicated significant at p < 0.001; **** indicated significant at p < 0.0001. Abbreviations: C4M, C8M, C1H, and C2H denote cold brewing for4 min, 8 min, 1 h, and 2 h, respectively; U4M, U8M, U1H, and U2H denote UACB for 4 min, 8 min, 1 h, and 2 h, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) L-theanine exhibited a significantly higher release rate constant in the UACB group than in the cold brewing group, increasing from 0.1404 to 0.2955 according to the kinetic analyses (p < 0.05; Fig. S2). Furthermore, its concentration in the UACB group remained significantly higher than that in the cold brewing group throughout the 2 h extraction period (p < 0.05; Fig. 6). After 8 min of UACB, its concentration reached 1.55 fold that of same duration cold brewing, approaching levels achieved after 1 h of cold brewing. This result is consistent with the sensory evaluation findings, indicating that the umami intensity of UACB tea after 8 min was close to that of tea infusion after 1 h of cold brewing. The key bitter-taste compounds, such as EGCG, ECG, and caffeine, showed increasing release rate constants in both UACB and cold brewing environments. Notably, the rate constant of caffeine was significantly enhanced under ultrasonic-assisted conditions (from 0.0985 to 0.1879; p < 0.05). As shown in Fig. 6, the concentration of caffeine after 8 min of UACB was 24.21 μg/mL, close to the concentration of 25.24 μg/mL obtained after 2 h of cold brewing. The extraction kinetics of non-galloylated catechins (e.g., EGC and EC) exhibited concentration change patterns similar to those of galloylated catechins (e.g., EGCG, ECG) (Fig. S2). The rapid extraction of these compounds could provide a sweet aftertaste to the tea infusion obtained from short-term ultrasonic-assisted brewing (Zhang et al., 2016). This allowed the bitterness profile of the tea infusion from 8 min of UACB to approach that of tea infusion from more than 1 h of cold brewing. However, prolonged ultrasonic extraction could cause severe damage to tea leaves, resulting in excessive extraction of bitter compounds (e.g., caffeine, EGCG). Previous studies have confirmed that high intensity of bitterness in green tea is a key factor contributing to the decline of overall flavor perception (Yang, Wang, et al., 2025). Conclusion This study combined metabolomics, kinetic analysis, and sensory evaluation to demonstrate the chemical basis of flavor quality formation during UACB. Results showed that mechanical disruption and cavitation accelerated the fragmentation of tea leaves, enhancing the extraction rate of quality compounds. Metabolomics revealed that while the types of extracted compounds were similar to those from cold-brewed samples, their concentrations were higher in UACB. For aroma quality, the accumulation of alcohols and aldehydes after 8 min of ultrasonic extraction imparted a rich floral and bean-like aroma to the tea infusion. Regarding taste and color quality, quantitative and kinetic analysis indicated that the extraction efficiency of L-theanine (umami sweetness) and flavonoids and catechins (bitterness) was higher in UACB than in cold extraction. The extraction rates of L-theanine in UACB after 8 min were comparable to those from 1 h of cold brewing. This enabled the tea infusion from UACB to achieve a rich, layered taste and color quality in a shorter time. This study provided a feasible strategy for improving cold-brewed tea extraction efficiency and a chemical basis for UACB. This method could reduce energy consumption in cold-brewed tea production by shortening extraction time compared to conventional methods and provide objective evidence for processing techniques. CRediT authorship contribution statement Chen Yang: Writing – original draft, Methodology, Formal analysis, Data curation, Conceptualization. Minghui Xu: Data curation. Wei Jiang: Data curation. Jiaqi Wang: Data curation. Xu Zhong: Data curation. Kaikai Wei: Data curation. Guijie Chen: Formal analysis. Qianying Dai: Funding acquisition, Formal analysis. Chuanjian Cui: Writing – review & editing, Methodology, Formal analysis. Ruyan Hou: Writing – review & editing, Funding acquisition, Formal analysis. Ethical approval All participants agreed and volunteered to participate in the complete sensory experiment after being fully informed of the study requirements and risks. The information of all participants participating in the sensory evaluation was protected, and the release of all sensory data was authorized by the participants. The sample raw materials required for this project are in line with food quality and safety, which will not cause harm to people, animals and the environment. There is no conflict of interest. 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.