Metabolomics Reveals How Scenting Reshapes the Chemistry of Jinhua Jasmine Tea

Jasmine tea owes its signature fragrance to an unusual manufacturing step: finished green tea is repeatedly mixed with fresh jasmine flowers so that the blossoms’ volatile compounds are adsorbed and retained by the leaves. A new study has now tracked this transformation in unusual chemical detail, following Jinhua jasmine tea from Zhejiang Province, China, through six processing stages, from the original baked green base tea through four successive scenting rounds to the final aroma-enhancement step known as Tihua. By combining non-targeted metabolomics, targeted quantification, electronic tongue and electronic nose measurements, and volatile profiling, the researchers showed that scenting progressively reshapes both the taste-related and aroma-related chemistry of the tea, with soluble sugars, amino acids and phenolic acids accumulating while catechins and flavonols generally declined.

Jasmine green tea is classified as a reprocessed tea and is typically produced by scenting baked green tea made from fresh leaves of Camellia sinensis with fresh flowers of Jasminum sambac, the double-petaled jasmine used in this work. The final sensory quality is jointly shaped by repeated tea-flower contact, heat dissipation and drying, and the fresh floral aroma and mellow yet brisk taste it imparts contribute to the high commercial value of jasmine tea among reprocessed teas in China. Previous research on jasmine tea processing has largely concentrated on the major production areas of Fujian and Guangxi, and although Jinhua jasmine tea is a representative product of the Jiangnan tea-producing region, its chemical changes across successive scenting stages had remained insufficiently characterized.

The production process followed in the study used baked green tea of the cultivar Longjing 43 as the base tea. For every 100 kilograms of base tea, 38, 30, 26 and 20 kilograms of fresh jasmine flowers were used in the first, second, third and fourth scenting rounds, respectively. Each round lasted 12 hours at approximately 35 degrees Celsius, during which the tea and flowers were mixed and piled, with intermittent spreading and turning to dissipate accumulated heat. At the end of each round the flowers were removed by sieving, and the tea was dried for one hour at 110 degrees Celsius after the first round and at 100 degrees Celsius after each of the remaining three rounds. Tihua then mixed 6 kilograms of flowers per 100 kilograms of base tea for 6 hours at about 35 degrees Celsius, with the flowers removed by sieving and no further drying. In total, 120 kilograms of fresh flowers were used per 100 kilograms of base tea, and three independent production batches provided three biological replicates per stage, 18 samples in all.

Instrumental sensory analysis revealed distinct stage-dependent patterns. On the electronic tongue, the scented samples generally showed higher responses than the base tea for sweetness, richness, umami, bitter aftertaste and astringent aftertaste, while sourness decreased after scenting. The authors emphasize that electronic tongue analysis provides instrumental taste responses rather than direct sensory perception, so these results should be read as objective indicators of taste variation during processing. The electronic nose, equipped with ten metal oxide sensors, likewise distinguished all six stages. The W1W sensor, which showed the highest overall response, rose sharply after the first scenting round and peaked there, whereas the responses of W1S and W2W continued to increase and reached their highest levels in the finished tea. Because these sensors have broad and overlapping response ranges, changes in individual sensor signals cannot be attributed to specific volatile compounds.

Non-targeted metabolomic analysis using ultra-high-performance liquid chromatography coupled with Orbitrap mass spectrometry annotated 1449 non-volatile metabolites across the six stages. Principal component analysis clearly separated the base tea from the scented samples along the first principal component, which explained 66.03 percent of the total variance, while the scented samples were arranged according to scenting stage along the second, indicating progressive chemical change. Heterocyclic and other compounds formed the largest category at 32 percent, followed by amino acids and organic acids at 16 percent, phenolic acids and benzenoids at 14 percent, lipids and terpenoids and flavonoids and derivatives at 12 percent each, carbohydrates, glycosides and organic oxygen compounds at 10 percent, and nucleotides and alkaloids at 4 percent. Comparing each scented stage with the base tea, 156, 170, 186, 236 and 267 metabolites were more abundant in the base tea for the successive comparisons, while 276, 322, 337, 346 and 324 were more abundant in the scented samples. Clustering of the differential metabolites highlighted sustained directional patterns, and pathway mapping placed them mainly in metabolic pathways, biosynthesis of secondary metabolites and biosynthesis of cofactors, with more specific annotations including flavone and flavonol biosynthesis, anthocyanin biosynthesis and amino acid biosynthesis.

Targeted quantification of 78 compounds, using authentic standards and triple quadrupole mass spectrometry, sharpened the picture of taste chemistry. The summed content of 13 soluble sugars increased significantly from the base tea to the finished tea, reaching its highest level at the final stage. Fructose, arabinose, galactose and mannose generally increased during scenting, whereas glucose peaked after the second round and then declined, and ribose, xylulose, glucuronic acid and xylose fluctuated. The summed content of 16 amino acids also rose significantly, again peaking in the finished tea. Grouped by reported taste attributes, the umami amino acids theanine, glutamic acid and aspartic acid increased significantly across the process, sweet amino acids rose with higher levels in the later stages, and bitter amino acids fell from the base tea to the first scented sample before increasing significantly thereafter. The first scenting round therefore shifted the balance among amino acid taste groups, while later stages were characterized more by overall accumulation.

The polyphenolic compounds told a different story. Among the 49 quantified polyphenols, the summed catechin content, despite a temporary increase after the second round, decreased significantly from the base tea to the finished tea, and the summed content of flavonols and flavones declined after the first round and remained below the base-tea level in all scented samples. In contrast, the summed phenolic acid content increased significantly and peaked in the finished tea. The authors suggest that the reduced catechin content may be associated with oxidation, hydrolysis, thermal transformation or other reactions caused by repeated scenting and drying, and that the rise in phenolic acids may partly reflect catechin transformation, since ester catechins such as EGCG, GCG and ECG can yield gallic acid and degalloylated flavanols during processing. Increases in ferulic acid, salicylic acid, p-coumaric acid and 4-hydroxybenzoic acid may also arise from the release or formation of low-molecular-weight phenolic compounds during repeated tea-flower contact and drying. The first round, which was followed by drying at the higher temperature of 110 degrees Celsius, marked a clear early redistribution among the three polyphenolic groups.

Correlation analysis between the 78 quantified compounds and the electronic tongue responses, restricted to significant positive correlations with p below 0.05 and r above 0.8, linked specific metabolites to instrumental taste signals. Epicatechin correlated positively with astringent aftertaste, bitter aftertaste, astringency, richness and umami. Several phenolic acids, including chlorogenic acid, ferulic acid, salicylic acid, gallic acid and vanillic acid, were positively correlated with sourness, aftertaste responses, astringency, richness, sweetness or umami. Theanine correlated with astringency, richness, sweetness and umami, while glutamic and aspartic acids correlated with richness and umami, and the sugars fructose, maltose and mannose correlated with umami, richness or sweetness. The authors caution that because both compound contents and electronic tongue responses varied across processing stages, these correlations may partly reflect common stage-dependent changes and should not be interpreted as evidence of direct sensory contributions.

Volatile profiling by headspace solid-phase microextraction coupled with gas chromatography-mass spectrometry detected 62 volatile compounds, whose summed content increased stepwise across the scenting stages and peaked in the finished tea. Esters and terpenes were the most abundant classes, followed by heterocyclic compounds and alcohols. The increase from the fourth scented sample to the finished tea is notable because Tihua involved a shorter final flower-tea contact and no drying step, which previous work suggests may favor the retention of some volatiles. Among 36 volatile compounds consistently differential in all comparisons with the base tea, most were present at lower levels before scenting and increased in the later stages. In the comparison of base tea and finished tea, ten additional differential compounds had odor activity values above 1, nine of them higher in the finished tea, including phenylethyl alcohol, ocimene, benzaldehyde, methyl jasmonate and (E)-beta-farnesene, which are associated with floral, fruity and sweet notes, while (Z)-3-hexenyl 2-methylbutanoate, described as green and herbal, was higher in the base tea.

Because content alone cannot capture aroma relevance, the researchers calculated odor activity values, dividing each compound’s content by its odor threshold in water. Of the 36 shared differential volatiles, 23 exceeded an odor activity value of 1 in at least one sample, and 15 surpassed a value of 50 in the finished tea, an operational cutoff rather than an established sensory threshold. These potential aroma contributors included methyl benzoate, delta-cadinene, (Z)-3-hexenyl acetate, linalool, methyl anthranilate, indole, alpha-farnesene, methyl salicylate, eugenol, benzyl acetate, (Z)-3-hexenyl benzoate, alpha-copaene, benzyl alcohol, hexyl benzoate and benzyl benzoate. Their trajectories were not uniform: linalool, methyl anthranilate, indole, alpha-farnesene and benzyl acetate continued to rise toward the finished tea, whereas methyl benzoate and (Z)-3-hexenyl acetate peaked earlier and declined. Several of these compounds, including linalool, benzyl acetate, alpha-farnesene, (Z)-3-hexenyl benzoate, methyl anthranilate and indole, had previously been reported as major aroma constituents of Jinhua jasmine tea. Taken together, the stage-resolved analysis portrays the first scenting round as an important early phase of chemical adjustment and Tihua as a final enrichment stage, with the highest summed contents of soluble sugars, amino acids and volatiles alongside selective changes in the odor activity profiles of potential aroma contributors.

Subject of Research: Chemical changes in non-volatile metabolites and volatile compounds during the scenting process of Jinhua jasmine tea

Article Title: Integrated non-targeted metabolomics, volatile profiling and targeted quantification characterize chemical changes in Jinhua jasmine tea during the scenting process

Article References: Zhang, Y., Li, M., Liu, S., Wang, J., Fu, D., Zhu, M., Liu, Z., & Wang, K. (2026). Integrated non-targeted metabolomics, volatile profiling and targeted quantification characterize chemical changes in Jinhua jasmine tea during the scenting process. Food Chemistry: X, 40, Article 104668. https://doi.org/10.1016/j.fochx.2026.104668

Image Credits: AI Generated

DOI: 10.1016/j.fochx.2026.104668

Keywords: jasmine tea, scenting process, metabolomics, volatile compounds, electronic tongue, electronic nose, catechins, amino acids, soluble sugars, odor activity value, Tihua, Jinhua