Effects of different types of water on the sensory and physicochemical properties of cold - brewed green tea

Abstract

Brewing water significantly influences tea infusion quality, yet its role in cold-brewed green tea remains underexplored. This study investigated the effects of four water types—direct drinking water (DDW), pure water (PW), mineral water (MW), and distilled water (DW)—on the sensory and physicochemical properties of cold-brewed green tea. Mineral dissolution varied by water type, with Se, Ca, Hg, Pb, and As being key differentiating elements. Brewing with MW resulted in a flat taste and dark color attributable to low flavor components and interactions involving Ca²⁺ and Mg²⁺. PW and DW led to higher catechin and caffeine levels, which intensified astringency and bitterness. In contrast, DDW enhanced sweetness, reduced bitterness, possessed a vibrant infusion color, and achieved the highest Se content. With its moderate Ca/Mg levels and neutral pH, DDW is optimal choice for cold brewing, offering practical guidance for both consumers and the tea industry.

Highlights

• •Brewing water composition significantly affects cold-brewed green tea quality.
• •Ca²⁺ and Mg²⁺ chelate galloylated catechins, thereby reducing bitterness.
• •DDW (moderate Ca²⁺/Mg²⁺, neutral pH) enhances sweetness and infusion color.
• •Selenium dissolution is highest in DDW, supporting its nutritional advantage. Cold brewing preserves heat-sensitive antioxidants better than hot brewing.
• •Cold brewing preserves heat-sensitive antioxidants better than hot brewing.

1. Introduction

Tea (Camellia sinensis (L.) Kuntze) is one of the most widely consumed beverages worldwide (Liu et al., 2025). Through millennia of cultural and technological evolution, diverse brewing methods have been developed, resulting in a rich variety of flavor profiles. Among these, cold brewing—prolonged low-temperature extraction—has gained increasing attention for producing infusions with reduced bitterness and enhanced sweetness (H.-L. Tan et al., 2023). Green tea—a non-fermented type—is particularly suited for cold brewing due to its delicate flavor profile and lingering-sweet aftertaste (Xue et al., 2023). In addition to its sensory attributes, green tea is a rich source of bioactive compounds, including polyphenols (notably catechins), caffeine, amino acids and essential minerals. These components produce a variety of health-promoting effects, such as potent antioxidant activity, anti-tumor potential, and protective benefits against cardiovascular and cerebrovascular disorders (Jiao et al., 2024; Lorenzo & Munekata, 2016).

The quality of tea infusion is governed by multiple factors, including tea cultivar, processing techniques, and brewing parameters (Thakur et al., 2025). As the final critical step prior to consumption, the brewing process facilitates the transfer of tea inclusions from leaves into water, ultimately forming a beverage with distinct sensory characteristics. During this process, the extraction efficiency of flavor-related and bioactive compounds is significantly modulated by water quality parameters (Cao et al., 2021). Substantial evidence has established that in hot brewing, ions such as Ca²⁺ influence the taste of green tea infusions by affecting polyphenols, proteins and organic acid extraction (Xu et al., 2013; H. Zhang et al., 2017), while water pH and salinity affect the color stability of water-soluble pigments including flavonols, anthocyanins, and flavanones (Zeng et al., 2017).

However, these findings are almost exclusively derived from hot-brewing conditions of 80–100 °C for 3–5 min (Li, Feng, et al., 2023). While previous studies have compared cold versus hot brewing using the same water type (Zannou et al., 2020),or examined the effects of different waters on hot-brewed tea (Muller et al., 2020), no study has systematically investigated how different water types affect the sensory and physicochemical properties of cold-brewed green tea. This knowledge gap motivated the present investigation.

This study focused on Yuhua tea, one of China's top ten famous tea products, to systematically investigate the impact of water quality on cold-brewed green tea. Four commercially available waters with distinct mineral compositions—DDW, PW, MW, and DW—were selected. Using a multi-platform analytical approach, this study compares the dissolution rates of minerals and bioactive compounds, evaluates the corresponding differences in sensory attributes and antioxidant capacity, and identifies the most suitable water type for cold-brewed green tea. The findings provide a theoretical foundation for quality control and optimization of cold-brewed tea products.

2. Materials and methods

2.1. Materials

2.1.1. Tea and water samples

The green tea samples (special grade), obtained from Fu Tong Tea Industry Co., Ltd. (Jiangsu, China), were processed from the same batch materials according to the traditional manufacturing technology of yuhua green tea, which includes withering, fixation, rolling and drying. Four types of water were used for brewing: direct drinking water (DDW), pure water (PW), mineral water (MW), and distilled water (DW). DDW was prepared from laboratory tap water using a reverse osmosis water purification system (ES-400G, Aidiwei Environmental Technology Co., Ltd., Zhejiang, China). PW was purchased from Yibao Co., Ltd. (Nanjing, China), MW from Nongfu Shanquan Co., Ltd. (Hangzhou, China), and DW from Watsons Food & Beverage Co., Ltd. (Guangzhou, China). All water samples were analyzed for pH, conductivity, and mineral composition prior to brewing, and DDW was freshly prepared on each day of experiment to ensure consistency.

2.1.2. Chemicals and reagents

Analytical chemical reagents, including calcium chloride (Sinopharm Chemical Reagents Co., Ltd., Shanghai, China), sodium chloride (Sinopharm Chemical Reagents Co., Ltd., Shanghai, China), ferric chloride (McLean Biochemical Technology Co., Ltd., Shanghai, China), disodium hydrogen phosphate (Aladdin Co., Ltd., Shanghai, China) and potassium dihydrogen phosphate (McLean Biochemical Technology Co., Ltd., Shanghai, China), were used in this study. Acetonitrile, formic acid, and methanol were all HPLC grade and from Merck (Darmstadt, Germany). Authentic standards were purchased from Shanghai Yuanye Biotechnology Co., Ltd. (Shanghai, China), including caffeine, gallic acid, (+)-Catechin (C), (−)-epicatechin (EC), (−)-gallocatechin (GC), (−)-epigallocatechin (EGC), (−)-catechin gallate (CG), (−)-epicatechin gallate (ECG), (−)-gallocatechin gallate (GCG), (−)-epigallocatechin gallate (EGCG). Dried NaOH was purchased from Shanghai Suke Chemical Co. Ltd. (Shanghai, China).

2.2. Preparation of tea infusions

The tea infusion was prepared according to the ratio of tea to water 1:50 (w/w). Cold brewing was conducted by adding water at room temperature (25 ± 2 °C) to the tea leaves in a wide-necked Schott bottle, followed by steeping in a refrigerator at 4 °C for 6 h. All other procedures were performed at room temperature unless otherwise specified. Each infusion was prepared in triplicate on separate days using independent biological replicates (separate batches of tea leaves and fresh water preparations). The same approach was applied to all subsequent experiments. The final tea samples were labeled as DDWT, DWT, MWT, and PWT according to the water type used.

2.3. Analysis of pH, conductivity and mineral content

The pH values were determined using a pH meter (SG2, Mettler-Toledo Instruments Co., Ltd., Shanghai, China). Buffer solutions at pH 4.01 and 7.01 (Mettler-Toledo) were used to calibrate the pH meter. Conductivities were measured using a conductivity meter (DDB-303, Shanghai Mai Instrument and Meter Co.,Ltd., Shanghai, China).

The mineral contents were determined in tea infusions applying a direct method: infusions were acidified with HNO3 to obtain a 0.2% (v/v) acid concentration and directly analyzed using ISIS-DS system. For tea leaves, samples were digested using closed microwave decomposition at 170 °C for 32 min following previously described procedures (Y. Ye et al., 2023). The element K was determined by the FAAS method. Other mineral elements, including Mg, Ca, Hg, As, Pb, Cd, Cr, Cu, Fe, Mn, Se and Zn, were determined using an ICP-MS (7700×, Agilent Technologies, Tokyo, Japan) operating with the follow conditions: RF power = 1150 W; Ar flow rate = 15 L min⁻¹; Ar auxiliary flow rate = 1 L min⁻¹; He flow rate = 5 and 10 mL min⁻¹; double pass spray chamber; micromist nebulizer with gas flow rate = 1.1 L min⁻¹ and 0.3–1.0 s for integration time (Peng et al., 2021). The dissolution rate of each metal was calculated using the following formula:

Metal dissolution ratio (%) =$\frac{A-B}{C\times 0.05}$ × 100%.

where A represents the metal content in tea infusion (μg/L), B represents the metal content in water (μg/L), C represents the metal content in green tea leaves (μg g⁻¹). 0.05 represents the brewing water volume (L).

2.4. Sensory evaluation

2.4.1. Taste analysis using electronic tongue

The taste profile of green tea infusion was measured using an Insent electronic tongue system SA402B (Insent Inc., Atsugi-chi, Japan). The data acquisition sequence was carried out alternately for the calibration solution and the tea infusion to be tested. Prior to data acquisition, the e-tongue was activated and calibrated to ensure data stability and reliability (Y. Ye et al., 2024). For testing, 35 mL of the tea infusion at 4 °C was transferred to an e-tongue cup, and three tests were conducted on each sample for subsequent analysis.

2.4.2. Sensory evaluation

Sensory evaluation was conducted by a trained panel of five assessors (three males and two females, aged 24–42 years) from the Tea Research Center, Jiangsu Academy of Agricultural Sciences. All assessors had prior experience in tea sensory evaluation. The evaluation was conducted in a standard tea assessment room with controlled environment (temperature 20–25 °C, humidity 40%–60%), neutral lighting and background colors, and proper ventilation to ensure no interference from external noise or odors. Standard tea evaluation method was used, and evaluators followed protocols to minimize personal influence on the results.

Tea infusions were prepared as described in section 2.2 and served at 4 °C in randomly coded cups (20 mL each). Assessors evaluated liquor color, aroma, and taste using a descriptive analysis method. Attributes were described using consensus vocabulary.

To ensure compliance with ethical standards, all members of the sensory evaluation panel were thoroughly informed about the study's procedures. Throughout the entire research process, strict protocols were implemented to protect the privacy of all relevant individuals. Ethical permission to conduct a human sensory study was granted by our institution. All participants provided written informed consent before the sensory evaluation. The consent form explained the confidentiality of responses and the voluntary nature of participation, and only those who gave an affirmative response proceeded with the evaluation.

2.4.3. Color measurement

Chromatic parameters of tea infusions were measured using a colorimeter (CM-23D, Konica Minolta Sensing Inc., Tokyo, Japan) with a D65 light source and 10° observer, using a cuvette (size: 50 × 38 mm; optical path: 10 mm)(Mao et al., 2021). The CIE L*a*b* system was adopted to describe the color of tea infusions, where L* stands for lightness ranging from −100 (dark) to 100 (white), a* represents the red (+)/green (−) color balance, and b* represents the yellow (+)/blue (−) color balance. C* (chroma) is an index of color saturation, with higher values indicating deeper color.

2.5. Physicochemical analysis of tea infusions

2.5.1. Analysis of catechins and caffeine

To determine the concentrations of catechins, gallic acid, and caffeine in the tea infusions, high-performance liquid chromatography (Agilent-1260 High Performance Liquid Chromatography, Germany) was utilized (Y. Ye et al., 2024). The samples were filtered through a 0.45 μm Millipore filter before injection and separated as follows: Agilent TC-C18 column (4.6 mm × 250 mm,5 μm; Agilent Technologies); column temperature 35 °C; post-run time 5 min; injection volume 5 μL; flow rate 0.8 mL min − 1; detection wavelength 278 nm. The mobile phases were A: 0.1% acetic acid; B: 100% acetonitrile. The elution solvent was initially 20% B, then ramped linearly to 25% B at 6 min, held at 30% B until 13 min, then ramped back to 20% B at 30 min.

2.5.2. Determination of scavenging capacity

The tea infusions were diluted to 50 μg/mL for in vitro antioxidant capacity analysis.

DPPH free radical scavenging rate was determined by using the corresponding kits. In brief, 400 μL diluted extracts and 600 μL DPPH (56 μg/mL) working solution were mixed in a 96-well plate (Deng et al., 2024). After 30 min, the absorbance (A₁) was measured at 517 nm. The DPPH scavenging activity was calculated using the following equation: DPPH• scavenging activity (%) =$(1-\frac{\mathrm{A}1-\mathrm{A}2}{\mathrm{A}0}$) × 100%, where A₁ is the sample absorbance, A₂ is the control (80% methanol instead of DPPH), and A₀ is the blank (80% methanol instead of sample).

ABTS assay was applied referencing the guidelines of the ABTS assay kit. In brief, 10 μL of diluted extracts or Trolox standard and 20 μL enzyme solution were mixed with 170 μL of ABTS working solution and reacted for 6 min. The absorbance was measured at 734 nm (Deng et al., 2024). For the control group, equal volumes of Trolox solutions of different concentrations were used instead of the sample solutions. For the blank group, equal volumes of distilled water were used instead of the sample solutions. A standard curve was prepared using Trolox solutions (0.1, 0.2, 0.4, 0.8, and 1.0 mM). The sample concentration was calculated from the standard curve.

Superoxide anion (O₂⁻•) was applied referencing the guidelines of the Superoxide anion kit (Y. Tan et al., 2025). In brief, 25 μL diluted extracts (A₁) or water (A₂) were mixed with 50 μL phosphate buffer (50 mmol L⁻¹, pH 7.8) and reacted for 1 min. Then, 1 mmol L⁻¹ hydroxylamine hydrochloride was added and incubated at 37 °C for 30 min, followed by 50 μL p-aminobenzenesulfonic acid (17 mmol L⁻¹) and 50 μL α-naphthylamine (7 mmol L⁻¹). After incubation at 37 °C for 20 min, absorbance was measured at 530 nm. The scavenging activity was calculated as: Superoxide anion scavenging activity (%) =$\frac{\mathrm{A}2-\mathrm{A}1}{\mathrm{A}2}$ × 100%.

Hydroxyl radical (•OH) was applied referencing the guidelines of the hydroxyl radical (•OH) kit (Jayabalan et al., 2008). The reaction mixture (900 μL) containing ferrous sulfate (6 μM), salicylic acid (18 μM), and H₂O₂ (3 μM) in potassium phosphate buffer (150 mM, pH 7.4) was incubated at 37 °C for 20 min. Then, 200 μL was transferred to a 96-well plate and absorbance was measured at 510 nm (A₁). For the control group, equal volumes of water were used instead of the sample solutions, recorded as A₂. For the blank group, equal volumes of water were used instead of H₂O₂ and sample, recorded as A₀. The scavenging activity was calculated as: Hydroxyl radical scavenging activity (%) =$\frac{\mathrm{A}2-\mathrm{A}1}{\mathrm{A}2-\mathrm{A}0}$ × 100%.

2.6. Statistical analysis

All results were recorded as mean ± standard deviation (SD) from three independent replicates. Statistical analysis was performed using one-way analysis of variance (ANOVA). When a significant difference was detected (p < 0.05), the Duncan test was used for post hoc multiple comparisons. Analyses were performed using SPSS statistics (IBM; Chicago, IL, USA). Principal component analysis (PCA) and Pearson correlation analysis were performed using Origin 2024 software (OriginLab, Northampton, MA, USA). For PCA, the data matrix consisted of 12 observations (4 water types × 3 replicates) and 13 mineral variables, and components with eigenvalues >1 were retained. For Pearson correlation, normality was confirmed by the Shapiro–Wilk test (p > 0.05 for all variables), and no extreme outliers were detected via boxplot inspection; the correlation matrix was based on 12 paired observations (4 water types × 3 replicates). Figures were generated using Origin 2024 and SPSS Graphics.

3. Results and discussion

3.1. Physicochemical properties of four water samples

The physicochemical properties of four water samples, including their pH, conductivity, and mineral contents, were shown in Table S1. The pH values of these water samples exhibited significant variation (p < 0.05). Specifically, the pH values of MW (7.74) and DDW (7.01) were above 7.0, while those of DW (5.77) and PW (6.66) were below 7.0. Similarly, the conductivity of the four samples also showed significant differences (p < 0.05). Notably, the conductivity of MW (71.48 μS/cm) was substantially higher than that of the other samples (46.66, 1.97, and 1.46 μS/cm for DDW, PW, and DW, respectively). Furthermore, a significant correlation was observed between pH value and conductivity in different types of brewing water (r = 0.890, p < 0.001). Overall, DW had the lowest pH and conductivity, while MW had the highest pH and conductivity. In addition to pH and conductivity, the mineral concentrations in the water samples also varied considerably (p < 0.05). Among them, K (20,000.00 ± 0 μg/L), Ca (3908.07 ± 79.49 μg/L) and Mg (1932.40 ± 17.67 μg/L) showed the highest contents among the four samples (Table S1). The concentration of the major elements (K, Ca, Mg) in MW was significantly higher than that in other water samples. To further explore these differences, principal component analysis (PCA) was performed using pH, conductivity, and mineral contents as raw data. The first two principal components (PCs) accounted for 54.3% and 26% of the total variance, respectively, effectively capturing the structural differences in mineral composition among the water samples (Fig. 1 A). Loading plot analysis suggested that Fe, Cu, Zn, and Mn might be the primary variables distinguishing the water samples. These findings align with previously reported characteristics of brewing water (Ma et al., 2024). The high mineral contents in MW and DDW also contributed to their high conductivity, which may in turn affect the flavor and quality of tea infusion (Zhang et al., 2024).

Fig. 1.

PCA and Person correlation analysis. A. PCA loading plot for water-related data. B. PCA loading plot for tea infusion data. C. PCA biplot. D. PCA score scatter plot. E. Importance dot plot of the random forest. DDW: direct drinking water; DW: distilled water; MW: mineral water; PW: pure water; DDWT: direct drinking water tea; DWT: distilled water tea; MWT: mineral water tea; PWT: pure water tea.

3.2. Effects of four water samples on dissolution of mineral substance of cold-brewed tea

The dissolution rate of minerals from tea leaves into infusion is a sensitive indicator of water-tea interaction, as it reflects the combined influence of water chemistry (e.g., pH, ionic strength, and mineral composition) on leaching efficiency (Bai et al., 2023). The concentrations of three major elements (K, Mg, and Ca), five heavy metals (Cr, As, Cd, Hg, and Pb), and five trace elements (Mn, Cu, Zn, Fe, and Se) in green tea leaves and tea infusions were determined by ICP-MS (Table S2). Among the major elements, K (20,340 ± 75.498 mg/kg), Ca (1953.744 ± 107.587 mg/kg), and Mg (1588.907 ± 76.720 mg/kg) showed the highest concentrations in green tea leaves. The dissolution rates of each element were calculated accordingly (Table S3). Results showed that dissolution rates varied considerably among the four water types. For example, the dissolution rate of K exceeded 65% across all groups, whereas that of Pb was consistently below 0.5%. Notably, MWT exhibited significantly higher dissolution rates for Ca and Cu than other infusions (p < 0.05, ANOVA with Duncan's test).

Principal component analysis (PCA) was performed on the dissolution rate data of the 13 elements. The PCA loading plot and biplot (Fig. 1B, 1C) revealed clear separation among the four tea infusion groups: (i) MWT was clearly separated from the other groups in the first quadrant (PC1+/PC2+); (ii) DDWT and PWT were primarily distributed in the negative PC2 region (PC2 < 0); and (iii) DWT in the negative PC1 region (PC1 < 0). These separations indicate that water type distinctly shapes the mineral dissolution profile. As shown in Fig. 1D, the first three principal components (PCs) accounted for 30.2%, 23.5%, and 15.8% of the total variance, respectively, cumulatively explaining 69.5% of variance.

The loading matrix (Table S4) was used to interpret which minerals drove these separations. PC1 (30.2%) was strongly positively associated with Ca (0.81), Mn (0.81), Cr (0.77), and Cu (0.64), and negatively associated with Hg (−0.71). PC2 (23.5%) was dominated by As (−0.85), Pb (0.72), and Se (−0.64). PC3 (15.8%) was mainly characterized by Cd (0.80) and K (0.66). Consistent with these loading patterns, the biplot (Fig. 1D) visually confirms that MWTI samples (first quadrant, PC1+/PC2+) are enriched in Ca, Mg, and Cu; DDWTI and PWTI (negative PC2 region) are associated with higher Se and As; and DWTI (negative PC1 region) with Hg. These patterns align with the dissolution rate data (Table S3).

Random forest analysis further identified the most discriminating elements. As shown in the importance dot plot (Fig. 1E), the key elements contributing to group differences were, in descending order: Se, Pb, Hg, Fe, As and Ca. All six elements exhibited absolute loading values >0.6 in the loading matrix (Table S4), demonstrating strong agreement between the two multivariate approaches. Notably, Se showed a significantly higher dissolution rate in DDWT than in other groups (Table S3). As an essential trace element, Se is a powerful antioxidant that can delay cell aging and promote immune cell proliferation (Minamoto et al., 2023), further supporting the potential nutritional advantage of using DDW for cold brewing. In this study, the contents of all detected heavy metals in green tea leaves were all below the limits specified in Chinese standard NY 659–2003 (i.e., 1 mg/kg for Cd, 0.3 mg/kg for Hg, 2 mg/kg for As, 5 mg/kg for Cr, 2 mg/kg for Pb in tea).

Given the limited sample size (n = 12), these PCA results should be interpreted as exploratory. Future studies with larger sample sizes are warranted.

While mineral dissolution profiles provide insight into water–tea interactions, sensory quality is also modulated by the complexation of minerals with polyphenols and other flavor compounds (Du et al., 2022). For example, Ca²⁺ in brewing water influences the quality of a tea infusion by inducing tea cream and sediment formation from combination of Ca²⁺ and organic acids (Xu et al., 2013). Al in the tea infusion mainly exists in the form of a complex with catechins. Although this restricts its bioavailability, it can improve the activation efficiency of umami receptors (Zhang et al., 2023). Therefore, the subsequent sections will integrate sensory and chemical analyses to determine the most suitable water for cold-brewed green tea.

3.3. Effects of four water samples on color of cold-brewed tea

Brewing water quality is one of the critical factors influencing the color characteristics of tea infusions. As shown in Table 1, the liquor color of cold-brewed green tea varied markedly with water type. DDWT exhibited a bright yellow-green appearance, while DWT appeared as clear light green. MWT showed a deep yellow, dull color, and PWT presented as light green with a yellowish tinge.

Table 1.

Sensory evaluation of cold-brewed green tea infusions prepared with four types of water.

Evaluate items	DDWT	DWT	MWT	PWT
Liquor color	Bright yellow-green	Clear light green	Deep yellow, dull	Light green with a yellowish tinge
Aroma	Clean & brisk	Fresh & brisk	Pure	Fresh & brisk
Taste	Fresh, brisk with a sweet aftertaste	Rich and mellow	Thin and plain	Brightly bitter

Note: DDWT: green tea infusions brewed with direct drinking water; DWT: green tea infusions brewed with distilled water; MWT: green tea infusions brewed with mineral water; PWT: green tea infusions brewed with pure water.

The study revealed significant differences in chromatic parameters among the four tea infusions (p < 0.05, Fig. 2A). Specifically, the L* values of green tea infusions cold-brewed with DW and PW were significantly higher than those brewed with DDW and MW, indicating a lighter color appearance.

Fig. 2.

The chromatic attributes of tea infusions cold brewed with four different water samples. A. The color of tea infusions cold brewed with four different water samples. B. Pearson correlation heatmap between water properties and tea infusion chromatic parameters. Correlation coefficients were calculated using 12 observations (4 water types × 3 replicates). Normality, linearity, and outliers were checked prior to analysis, and no major violations were found. * p < 0.01; ** p < 0.001.

The effects of brewing waters on the color of tea infusions appeared to be related to their pH and mineral contents, with a significant correlation between these factors and the chromatic parameters of the tea infusions (Fig. 2B). The L* was negatively correlated with pH (r = −0.92, p < 0.001) and the mineral contents, especially Ca (r = −0.86, p < 0.001) and Mg (r = −0.90, p < 0.001), indicating that high pH produced darker infusions. This is because catechins will be oxidized in water with higher pH value and form coloured substances (Long et al., 2024). In addition, the clarity of the tea infusion brewed with MW is significantly reduced. This is mainly because Ca²⁺ and other divalent metal ions promote the formation of tea polyphenol–metal complexes, thus affecting the light transmittance. (Ouyang et al., 2017).

The a* values of green tea infusions were negative across all samples. Notably, MWT exhibited a significantly lower a value than the other three groups, indicating a stronger green tone. The values of b* and C* decreased in the order MWT > DDWT > PWT > DWT, indicating that higher pH and conductivity intensified the yellowing of tea infusions. The C* was positively correlated with pH (r = 0.94, p < 0.001) and the mineral contents, especially Ca (r = 0.86, p < 0.001) and Mg (r = 0.84, p < 0.001). Consequently, when cold brewed with weakly acidic waters like DW and PW, green tea infusions exhibited duller coloration and lower saturation.

In summary, water quality systematically influences tea infusion color by modulating pH-dependent and ion-driven chemical interactions. MW (high pH) yields excessively dark infusions, while weakly acidic PW and DW result in undersaturated colors. For green tea—an unfermented tea best presented with lighter color—we recommend DDW (neutral pH, moderate Ca/Mg content) for cold brewing.

3.4. Effect of four water samples on taste quality of cold-brewed tea

The taste characteristics of cold-brewed green tea infusions were evaluated using sensory evaluation and electronic tongue (ET) analysis. Sensory evaluation is a traditional and widely used method for assessing tea taste, but it is inherently subjective and influenced by panelist experience and physiological differences. In contrast, the ET is a bionic detection system composed of cross-sensitive sensor arrays with varying functions. It can detect the overall characteristic signals of sample solutions and identify liquid-phase sample fingerprints using appropriate pattern recognition methods (Chen et al., 2025). In this study, the ET data were normalized due to significant variations in output data ranges across sensors, and the processed data were visualized as pie charts, where each segment represented a specific taste attribute visually illustrating the taste profile of each tea infusion.

The sensory evaluation and ET data showed strong correspondence. Sensory evaluation (Table 1) showed that PWT was described as “brightly bitter,” the pie chart (Fig. 3) showed the highest value of bitterness along with the lowest value of sweetness and umami. DDWT, in contrast, was described as “fresh and brisk with a sweet aftertaste,” and the ET showed the highest sweetness value and the lowest bitterness value among all samples (p < 0.05). In general, the umami and sweetness correlate positively with the grades and quality of green tea (J.-H. Ye et al., 2022). As for MWT and DWT, the sensory evaluation distinguished them as “thin and plain” versus “rich and mellow.” The ET revealed that DWT had a notably higher bitterness level than MWT, while no significant difference was observed in umami or astringency between the two. This suggests that the perceived “richness” of DWT may originate from its higher levels of bitter compounds (Table 1), rather than from umami or astringency. In terms of aroma, DDWT was rated as “clean & brisk,” DWT and PWT both as “fresh & brisk,” and MWT as “pure”.

Fig. 3.

Taste profiles of cold-brewed green tea infusions prepared with different types of water analyzed by electronic tongue. Each segment represents the relative intensity of bitterness, astringency, richness, saltiness, sweetness, and umami. Data were normalized for comparison. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

These taste differences can be attributed to the variation in mineral content among the four water types. MW contains abundant Ca²⁺ and Mg²⁺, which can chelate with galloylated catechins (e.g., EGCG and ECG)—known agonists of TAS2R bitter receptors—thereby reducing the free concentration of these bitter compounds and potentially hindering their binding to TAS2R39 and TAS2R14 (Narukawa et al., 2011; Zhang et al., 2020). This may explain the lowest bitterness score of MWT. Conversely, DW and PW, with minimal Ca²⁺/Mg²⁺, allow galloylated catechins to remain largely uncomplexed, resulting in higher bitterness in PWT and DWT. Regarding sweetness, DDWT exhibited the highest sweetness score (Fig. 3). Studies have shown that moderate levels of Ca²⁺/Mg²⁺ can reduce the bitterness of EGCG and caffeine solutions, while excessive Mg²⁺ suppresses both bitterness and sweetness perception (Xu et al., 2024). This aligns with our observations: DDW (moderate Ca²⁺/Mg²⁺) reduced bitterness without compromising sweet taste, whereas MW (excessive Ca²⁺/Mg²⁺) also suppressed sweetness, and PW/DW (extremely low Ca²⁺/Mg²⁺) failed to suppress bitterness, resulting in lower sweetness scores.

Thus, based on the overall taste profile, DDW is preferable for cold-brewed green tea, as it enhances sweetness while reducing bitterness compared to other water types.

3.5. Effect of four water samples on the main chemical composition of cold-brewed tea

The eight main catechins in green tea infusions can be categorized into epi- catechins (EC, EGC, ECG, EGCG) and non-epi-catechins (C, GC, CG, GCG), which are primarily responsible for the infusion's astringency and bitterness (Y. Ye et al., 2018).

In this study, the epi- catechins content (9.068 ± 0.196%–6.969 ± 0.507%, Table 2) was significantly higher than that of non-epi-catechins (0.543 ± 0.010%–0.418 ± 0.020%), largely due to EGCG (4.495 ± 0.100%–3.316 ± 0.255%), which had the highest concentration among all identified catechins. Previous studies indicate that EGCG accounts for approximately 40–50% of the total catechin content in green tea, making it the most abundant and significant catechin (J. Li et al., 2025). Additionally, water quality significantly influenced catechin extraction: cold-brewed tea prepared with MW had significantly lower epi-catechins content (p < 0.05) than that brewed with DW, PW, or DDW. However, no significant difference was observed in epicatechin content between tea cold-brewed with DW and PW. For non-epi-catechins extraction, significant variations were found among water types (p < 0.05). Extraction efficiency decreased in the following order: DWT (0.543 ± 0.010%) > PWT (0.524 ± 0.011%) > DDWT (0.496 ± 0.013%) > MWT (0.418 ± 0.020%).

Table 2.

Catechins content of cold brewing green tea infusion brewed by four water samples.

Content (%)	DDWT	DWT	MWT	PWT
GA	0.30 ± 0.01a	0.29 ± 0.00b	0.26 ± 0.02c	0.29 ± 0.01b
GC	0.35 ± 0.01b	0.37 ± 0.01a	0.29 ± 0.02c	0.37 ± 0.01a
EGC	2.11 ± 0.03b	2.21 ± 0.02a	1.86 ± 0.12c	2.30 ± 0.05a
C	0.15 ± 0.00b	0.17 ± 0.00a	0.13 ± 0.00c	0.16 ± 0.00b
EGCG	4.10 ± 0.05c	4.29 ± 0.01b	3.32 ± 0.26d	4.50 ± 0.10a
CAFE	3.69 ± 0.07b	3.75 ± 0.01a	3.26 ± 0.19c	3.79 ± 0.08a
EC	0.93 ± 0.02b	0.97 ± 0.03a	0.83 ± 0.05c	0.98 ± 0.03a
GCG	ND	ND	ND	ND
ECG	1.21 ± 0.05b	1.24 ± 0.01b	0.97 ± 0.09c	1.29 ± 0.02a
CG	ND	ND	ND	ND
Epi-catechins	8.34 ± 0.13b	8.70 ± 0.05a	6.97 ± 0.51c	9.07 ± 0.20a
Non-epi-catechins	0.50 ± 0.01b	0.54 ± 0.01a	0.42 ± 0.02c	0.52 ± 0.01a
Total catechins	12.82 ± 0.22b	13.29 ± 0.07b	10.91 ± 0.73c	13.67 ± 0.29a

Note: Data are presented as mean ± SD (n = 3). Different superscript letters (a–d) in the same row indicate significant differences among different water types (p < 0.05). ND: not detected.

The observed differences in catechin extraction under cold brewing conditions (4 °C, 6 h) can be explained by two potential mechanisms. First, complexation between Ca²⁺/Mg²⁺ and galloylated catechins may limit extraction (Li, Lu, et al., 2023). At low temperatures, the solubility of catechin–metal complexes is reduced, promoting precipitation and thereby reducing the concentration of free catechins in the infusion. This explains why MWT, brewed with MW containing the highest Ca²⁺ and Mg²⁺ concentrations (Table S1), exhibited the lowest catechin content. Second, pH-dependent degradation of catechins may also contribute, i.e., alkaline conditions accelerate catechin degradation (Y.-Q. Xu et al., 2019). Although both mechanisms may be involved, the lower catechin content in MWT is likely primarily attributable to reduced extraction efficiency due to complexation with Ca²⁺/Mg²⁺, given that the cold brewing temperature (4 °C) would substantially slow down degradation reactions.

Under cold brewing conditions, the extraction of bitter compounds (e.g., EGCG, caffeine) is reduced by over 40%, while umami compounds (e.g., L-theanine) are extracted more readily (Yang et al., 2025). Water quality can further modulate this pattern. Galloylated catechins (EGCG and ECG) are recognized as the primary contributors to bitterness and astringency, while non-galloylated catechins (EGC and EC) contribute to sweet aftertaste (Cao et al., 2019; Y.-Q. Xu et al., 2018). In the present study, DWT and PWT, brewed with DW and PW (low mineral content, weakly acidic pH), exhibited the highest levels of EGCG and ECG, explaining their pronounced bitterness and astringency. MWT, with the lowest levels of galloylated catechins, showed a “thin and plain” taste. DDWT, brewed with DDW (neutral pH, moderate Ca²⁺/Mg²⁺), achieved a favorable balance—moderate catechin extraction that avoided excessive bitterness while preserving sweetness and umami.

Correlation analysis demonstrated significant negative associations between water characteristics (pH and conductivity) and catechin content in tea infusions (Fig. 4). Three predominant catechins—EGCG, EGC, and EC—showed particularly strong inverse relationships with water conductivity (r = −0.89, −0.89, and − 0.86, respectively; p < 0.001). These negative correlations are meaningful because catechins are key determinants of tea taste. Galloylated catechins such as EGCG are primary contributors to bitterness and astringency, while non-galloylated catechins such as EC contribute to sweet aftertaste (Battestin et al., 2008; Cao et al., 2019; L. Zhang et al., 2020).

Fig. 4.

Pearson correlation heatmap between water quality parameters (pH, conductivity) and catechin components in cold-brewed green tea infusions. The color scale indicates correlation coefficients. * p < 0.05; ** p < 0.001. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

In summary, the mineral content of brewing water significantly influences both the extraction efficiency and the stability of catechins in cold-brewed green tea. These chemical differences, in turn, determine the bitterness, astringency, and sweet aftertaste of the final infusion.

3.6. Comparison of antioxidant capacity between cold-brewed and hot-brewed tea using DDW

Based on the findings above, DDW (neutral pH, moderate Ca²⁺/Mg²⁺) was identified as the most suitable water type for cold-brewed green tea. To further evaluate the differences between brewing methods, we compared the in vitro antioxidant capacity of green tea infusions prepared with DDW under cold brewing (4 °C, 6 h) and hot brewing (100 °C, 5 min).

As shown in Fig. 5, cold-brewed tea exhibited significantly higher scavenging activities against •OH, DPPH•, ABTS⁺, and O₂⁻• compared to hot-brewed tea. First, heat-sensitive bioactive compounds, particularly galloylated catechins (e.g., EGCG and ECG), are better preserved under low-temperature extraction. The high temperature (100 °C) during hot brewing accelerates oxidative degradation of these components by promoting oxidation reactions with oxygen (generating inactive quinones) and inducing polymerization into larger insoluble complexes (Yuan et al., 2024). In contrast, the low-temperature environment (4 °C) of cold brewing inhibits this process and preserves the structural integrity of tea polyphenols. Second, during cold brewing, the slower thermal movement of water molecules tends to extract antioxidants with moderate polarity (such as flavonoids and polyphenols), while avoiding excessive release of bitter substances (e.g., caffeine and theasaponin) that are easily dissolved during hot brewing (Cao et al., 2021). This selective extraction results in a higher concentration of effective antioxidant components in cold-brewed green tea, thereby enhancing its ability to scavenge free radicals. .

Fig. 5.

Antioxidant activity of green tea infusions prepared with DDW using cold (CBT) and hot (HBT) brewing methods. (A) •OH, (B) DPPH•, (C) ABTS•+, (D) O2 − •. Different lowercase letters (a, b) above the bars indicate significant differences between CBT and HBT for each assay (p < 0.05). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Nevertheless, it is important to note that in vitro radical scavenging assays measure chemical reactivity under controlled conditions and do not directly reflect in vivo health benefits (Menya et al., 2024). Factors such as absorption, metabolism, and bioavailability of tea polyphenols in the human body are not accounted for in these assays. Therefore, while cold-brewed green tea shows superior in vitro antioxidant activity compared to hot-brewed tea, this finding should not be overinterpreted as direct evidence of greater health benefits in a physiological context.

In summary, cold-brewed green tea prepared with DDW preserves heat-sensitive antioxidant components more effectively than hot-brewed tea, resulting in higher in vitro radical scavenging capacity. However, the physiological relevance of this difference remains to be established through future in vivo studies.

3.7. Limitations and future directions

This study used only one green tea product (Yuhua tea), which limits the generalizability of our findings. Future work should include a broader range of green tea products (e.g., Longjing, Biluochun) as well as other tea types (e.g., black tea, oolong tea, white tea) to validate the observed water-type effects. Furthermore, integrating tea physicochemical parameters with water quality indicators could enable the development of a predictive model for water–tea compatibility, providing personalized brewing recommendations for consumers and the tea industry.

4. Conclusions

The quality of cold-brewed green tea is significantly influenced by water mineral composition and pH. MW, with the highest Ca²⁺/Mg²⁺ and alkaline pH, yielded the darkest infusion color and the lowest catechin extraction, resulting in a flat taste. PW and DW, having low mineral content and acidic pH, led to higher catechin levels and intensified bitterness. In contrast, DDW (neutral pH, moderate Ca²⁺/Mg²⁺) reduced bitterness, preserved sweetness, and produced a vibrant yellow-green hue, making it the optimal choice for cold brewing. Additionally, Se—a beneficial trace element—showed higher dissolution in DDWT, highlighting its nutritional advantage. Notably, cold-brewed tea better preserves heat-sensitive antioxidants than hot-brewed tea, resulting in superior in vitro radical scavenging capacity. These findings provide practical guidance for consumers and the tea industry in selecting brewing water to optimize flavor and nutritional quality.

CRediT authorship contribution statement

Yutong Ye: Writing – original draft, Methodology, Investigation. Lingfei Ji: Software, Data curation. Yiyang Yang: Writing – review & editing, Project administration, Funding acquisition, Conceptualization.

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.