Topic hub — 2026-09-07
Tea's bitterness and astringency
Tea's bitterness and astringency are not decided in a single place. Catechins and caffeine are named as the carriers, and whether catechins or something else is the main contributor to astringency is not agreed between reports on green tea and reports on black tea. Shading and picking move the amount of these carriers themselves, depending on whether the leaf was shaded and when it was picked, and water temperature and steeping rewrite, cup by cup, how much of the carriers are drawn out of the same leaf. Firing and post-fermentation are more irreversible processes that reduce the carrier molecules or remake them into other molecules. And in Chinese tea tasting, a time axis called huigan is also discussed — sweetness that rises with a delay after the bitterness and astringency fade. Here, only the claims confirmed at each stage — leaf, water, and fire — are laid out, kept separate.
On this page (12 sections)
- What this page establishes
- The carriers — catechins, caffeine, and theanine
- What shading and picking move — matcha, tencha, and gyokuro
- What water temperature and steeping move
- What firing and roasting move
- What disappears with post-fermentation — goishicha, awa-bancha, and pu-erh
- Huigan, a time axis
- Frequently asked
- What is not known
- Read in the encyclopedia
- Go deeper in the research notes
- Source ledger
What this page establishes
- That the bitterness of green tea infusions correlates with the concentration of epigallocatechin gallate and epicatechin gallate has been confirmed in both independent sensory studies and receptor studies.
- While one report treats catechins as the main carrier of astringency in green tea, a report on black tea found a different glycoside to be dominant instead, and the two do not agree across the teas studied.
- Shading does not necessarily move bitterness and astringency in the same direction; one report finds that only the weakly bitter-astringent epigallocatechin decreases selectively.
- Huigan is widely discussed as part of tea culture, but no study has yet been found, for tea, that measured its rise as a continuous time curve.
The carriers — catechins, caffeine, and theanine
In a study examining the relationship between concentration and taste, the bitterness of green tea infusions correlated with the concentration of epigallocatechin gallate (EGCG) and epicatechin gallate (ECG).
Scope: Green tea infusions and the concentration-taste model in the cited study
Note: The correlation and fitted curve are specific to this study and are not a universal conversion formula.
Show data
ingredient-green-tea —has_bitterness_correlated_with→ "The concentration of EGCG and ECG"Reports are divided on which carrier is mainly responsible for astringency. The green tea literature treats catechins as the main contributor to astringency, while a taste-dilution study on black tea reported flavonol glycosides to be dominant instead — the two do not agree across differences in the tea studied and the method used.
Scope: Tea infusions. Asserted for green tea; contradicted for black tea.
Note: The green tea literature regards catechins as the main contributor to astringency, but a taste-dilution study on black tea found flavonol glycosides to be dominant instead. The matrix (green tea vs. black tea) and the method (correlation vs. dilution-to-threshold) differ.
Show data
compound-catechins —are_main_contributors_to→ sensation-astringentEven among catechins, the strength of bitterness varies with the molecule's shape. From the order in which catechins bearing a galloyl group show a stronger receptor response at the receptor hTAS2R39 than those without one, it has been inferred that galloylation strengthens bitterness.
Scope: The role of the galloyl group is inferred from the hTAS2R39 response order (ECg, EGCg >> EC, EGC), described as resembling human taste-intensity observations.
Note: A direct human dose-response is not part of this assay. The human comparison in the receptor study is a citation, not something the study itself carried out.
Show data
compound-catechins —have_bitterness_increased_by_galloylation→ sensation-bitterApart from the components that carry astringency and bitterness, the umami compound theanine acts directly on the bitterness receptor itself. In an experiment using cells expressing this receptor, L-theanine was reported to antagonistically block the bitter receptor TAS2R16.
Scope: The same hTAS2R16-expressing cell system
Note: Reported as partial, non-surmountable antagonism. Not extrapolated to the overall bitterness of tea or dishes.
Show data
compound-l-theanine —antagonizes→ receptor-tas2r16What shading and picking move — matcha, tencha, and gyokuro
Tencha, the raw material for matcha, is made from tea leaves grown under a cultivation method called shading cultivation (fukushita saibai), which involves blocking sunlight.
Note: Tencha's raw tea leaves are new shoots grown under shading cultivation (covered cultivation).
Show data
ingredient-tencha —requires→ process-tea-shading-cultivationShading does not necessarily move bitterness and astringency in the same direction. A source covering new shoots grown under high-shading-rate materials for roughly 20-30 days — as for gyokuro and tencha — lists the increase in caffeine under a "bitterness / functional" column and the decrease in catechins under a "bitter-astringent components" column, as separate items.
Scope: New shoots of tea plants -- such as for gyokuro and tencha -- grown under high-shading-rate materials for roughly 20-30 days
Note: A re-investigation of an item deferred in W44 (shading cultivation, "astringency only"). The existing claim covered only the change in catechins (astringent, bitter-astringent compounds) and did not state bitterness explicitly. This claim records that, within the same table of the same manual, the increase in caffeine is classified in the remarks column as "bitterness / functional," while the decrease in catechins is classified as "bitter-astringent compounds" -- a separate item -- showing that the source explicitly distinguishes bitterness from bitter-astringency (astringency).
Show data
process-tea-shading-cultivation —increases→ compound-caffeineIn the composition of matcha itself, too, shading is thought to act differently depending on the type of catechin. The catechins that fall with shading are uneven: it is the weakly bitter-astringent epigallocatechin (EGC) that falls, while the strongly bitter-astringent epigallocatechin gallate (EGCG) barely falls at all.
Scope: Matcha made from tencha (shading cultivation)
Note: This is background text in the "Purpose" paragraph of a NARO (National Agriculture and Food Research Organization) conference abstract, stated as a summary of prior knowledge rather than new measurement data from the abstract itself (citation numbers are not given in the abstract). Figures such as the rate of change in EGC and EGCG content are not given in this abstract.
Show data
ingredient-matcha —has_composition→ "The source states that the catechin whose content falls with shading cultivation is epigallocatechin (EGC), which has weak bitter-astringency, while epigallocatechin gallate (EGCG), which has strong bitter-astringency, shows almost no drop in content."The time elapsed after the shading is removed also moves the composition. In a field trial using the gyokuro cultivars 'Yabukita' and 'Okuyutaka' from Fukuoka Prefecture, lengthening the exposure to light after the shading was removed decreased free amino acids and theanine in the crude tea, while the catechin content (the sum of EC+ECG+EGC+EGCG) increased. In 'Yabukita,' theanine fell from 2224 mg/100g under shading to 1983 mg/100g after 6 hours of exposure, and catechins rose from 9.4% to 10.3% (dry weight).
Scope: Fukuoka Prefectural Agriculture and Forestry Research Center, Yame Branch; gyokuro cultivars 'Yabukita' and 'Okuyutaka'; a 2006 field trial (3 replicates)
Note: The subject is gyokuro, not tencha (the raw material for matcha) itself. However, gyokuro and tencha belong to the same family of shading-cultivation techniques and involve the same mechanism -- shading changing amino acid / catechin composition. Note that the sensory evaluation uses overall items such as "flavor" from the standard tasting method, not a scale that isolates bitterness alone.
Show data
ingredient-gyokuro —has_composition→ "The longer the exposure time (0/1/2/4/6 hours) after removing the shading material, the more the free amino acid and theanine content in the crude tea fell, while the catechin content (the sum of EC+ECG+EGC+EGCG) rose. In the 'Yabukita' cultivar, theanine fell from 2224 mg/100g under shading to 1983 mg/100g after 6 hours of exposure, and catechins rose from 9.4% to 10.3% (dry weight). Dimethyl sulfide (DMS), the main component of the covering aroma, fell to about 55% of the shaded level after 1 hour of exposure ('Yabukita') and to about 88% after 2 hours of exposure ('Okuyutaka'). Sensory evaluation by the standard tasting method (internal quality such as flavor) tended to worsen the longer the exposure time ('Yabukita' total difference: -0.7 points at 1 hour of exposure, -3.2 points at 6 hours)."What water temperature and steeping move
A study on Tieguanyin infusions reported that the tea-to-water ratio, water temperature, steeping time, and number of infusions all affected the combined bitterness-and-astringency attribute.
Scope: An orthogonal regression design and trained panel using Tieguanyin
Note: The evaluated attribute is the composite "bitterness-astringency." It does not claim to have varied bitterness and astringency separately.
Show data
ingredient-tieguanyin-infusion —has_bitter_astringency_affected_by→ "Tea-to-water ratio, water temperature, steeping time, and number of infusions"Water temperature can also remake the carrier molecules themselves into other molecules. In an experiment extracting Turkish green tea at 75, 85, and 95°C for 1-45 minutes, the main reason EGCG was lower at 95°C extraction than at 85°C was reported to be epimerization at high temperature into the non-epi-form gallocatechin gallate (GCG).
Scope: Turkish green tea extracted at 75/85/95 C for 1-45 minutes. GCG, GC, and C continued to increase with temperature and time (GCG reaching a maximum of 3.63 mg/100 ml at 95 C, 45 minutes).
Note: A single tea material (fresh Turkish green tea). The authors note that leaf age, size, season, and manufacture affect the results.
Show data
compound-egcg —epimerizes_to→ "Non-epi-form gallocatechin gallate (GCG) during high-temperature extraction. The main reason EGCG was lower at 95 C extraction than at 85 C was this epimerization."Not just how many times a tea is steeped, but how many times it is sipped, also moves how bitterness is felt. In an experiment with repeated sips of PROP solution, coffee, mate tea infusion, and grapefruit juice, the effect of each sip was reported to accumulate and push up the maximum bitterness intensity.
Scope: Multiple-sip time-intensity in normal-BMI and overweight groups
Note: The abstract reports that the effect of each sip was significant, and the cumulative effect was larger for samples with stronger bitterness.
Show data
process-repeated-sipping —can_increase→ "The maximum bitterness intensity of PROP solution, coffee, mate tea infusion, and grapefruit juice"What firing and roasting move
Roasting tea into hojicha reduces the catechins involved in astringency. In an analysis comparing crude teas of first flush, cut-leaf, and second flush from Kochi Prefecture across roasting degrees, total catechins (mg per 100 ml of tea liquor) decreased as roasting deepened: 34.89 to 21.86 to 14.59 to 6.98 for first flush, 25.31 to 12.09 to 8.82 to 2.33 for cut-leaf, and 54.51 to 26.31 to 18.26 to 8.68 for second flush.
Scope: Kochi Prefectural Agricultural Technology Center Tea Experiment Station; crude teas of first flush, cut-leaf, and second flush with a 50-second steaming time. Roasted 2 times with an automatic continuous hojicha roaster (exhaust temperature 90 C for the 1st roast, 100 C for the 2nd roast, with drum rotation speed adjusted for each flush and roasting degree).
Note: Direct quantification by HPLC (JASCO X-LC system). Only 1 sample per condition; the number of replicates is not stated.
Show data
process-hojicha-roasting —reduces→ compound-catechinsHow caffeine decreases is not as simple as catechins. In the analysis from the same region, light roast showed almost no decrease — first flush 2.61 mg/100ml (101% of crude tea), cut-leaf 2.50 (99%) — whereas deep roast decreased by 6-13%: first flush 2.43 (94%), cut-leaf 2.17 (87%).
Scope: Kochi Prefectural Agricultural Technology Center Tea Experiment Station, 2014; light-roast and deep-roast hojicha made from crude first-flush and cut-leaf tea (roasted 2 times with an automatic continuous hojicha roaster).
Note: Direct quantification by HPLC (Shimadzu).
Show data
process-hojicha-roasting —reduces→ compound-caffeineHow the temperature and time of firing move the sweetness of aroma and taste has also been traced experimentally. In a trial that fired grade-A crude tea at combinations of tea temperature 90-120°C and time 10-40 minutes, 90°C for 10 minutes gave an aroma score of 4, roast aroma 0, burnt smell 0, and sweetness of taste 0, whereas 120°C for 40 minutes gave an aroma score of 9, roast aroma 4, burnt smell 1, and sweetness of taste 3 — roast aroma and burnt smell intensified as temperature and time rose.
Scope: Grade-A crude tea (from Kochi Prefecture, 50-second steaming time).
Note: The infusion was prepared with 4g of tea leaves, water at 80 C, 140 ml of water, and an 80-second steeping time, then tasted. Aroma was evaluated by sensory assessment of the scent of 4g of tea leaves steeped in 140 ml of hot water. The number of evaluators is not stated.
Show data
process-tea-hiire —has_sensory_response→ "Using a small rotating-drum firing machine (with tea-temperature control) modified from a Collins roaster, grade-A crude tea was fired at combinations of tea temperature 90, 100, 110, and 120 C and time 10, 20, 30, and 40 minutes, and aroma (9-point scale, 5 being favorable) and the infusion (roast aroma 0-4, burnt smell 0-1, sweetness of aroma 0-3, sweetness of taste 0-3) were evaluated (2020). At 90 C for 10 minutes, the aroma score was 4, roast aroma 0, burnt smell 0, sweetness of aroma 0, and sweetness of taste 0, whereas at 120 C for 40 minutes the aroma score was 9, roast aroma 4, burnt smell 1, sweetness of aroma 2, and sweetness of taste 3 -- roast aroma and burnt smell intensified as temperature and time increased."What disappears with post-fermentation — goishicha, awa-bancha, and pu-erh
In teas that have gone through post-fermentation, cases have been reported where the catechins involved in astringency become markedly scarce. HPLC analysis of hot-water extracts of goishicha reported that the total amount of the main catechins and gallic acid was about 1/7 that of green tea, and about half that of oolong and black tea.
Scope: A comparison of hot-water extracts across several teas including goishicha
Note: As described by Ukeda (2015); the original is given as Moriyama, Katayama, Nakabayashi, Ukeda, and Sawamura, Nippon Shokuhin Kagaku Kogaku Kaishi 49, 679 (2002). The primary source (the 2002 paper) itself was not checked this time, so it is included as a secondary citation.
Show data
ingredient-goishi-cha —has_composition→ "HPLC analysis of hot-water extracts found the total amount of the main catechins and gallic acid in goishicha to be extremely low compared with other teas -- about 1/7 that of green tea and about half that of oolong and black tea, the source reports. Because catechins are the components involved in tea astringency, the source describes goishicha, whose catechins decrease during fermentation, as having mild astringency."Taste-sensor measurements also show low astringency for post-fermented teas. Compared by taste sensor, awa-bancha showed weaker astringency stimulus and umami than sencha and black tea, with only its sourness approaching the level of Ishizuchi-kurocha.
Scope: A comparison of infusions of awa-bancha, Ishizuchi-kurocha, sencha, and black tea (Tokushima Prefectural Industrial Technology Center, taste sensor SA402B)
Note: A comparison between single samples; the number of replicates and statistical testing are not stated.
Show data
ingredient-awa-bancha —has_sensory_attribute→ "Measuring infusions with a taste sensor (SA402B), awa-bancha showed lower output values for astringency stimulus and umami than sencha and black tea. Its sourness output value was the second highest, after Ishizuchi-kurocha. These results are reported to show that awa-bancha has a sour, clean-tasting profile."Leaf analysis points the same way. Tannin content is reported as 8.1 g/100g for awa-bancha, 7.5 g/100g for Ishizuchi-kurocha, 11.4 g/100g for sencha, and 6.5 g/100g for black tea; caffeine content as 1.5 g/100g for awa-bancha, 1.4 g/100g for Ishizuchi-kurocha, 2.1 g/100g for sencha, and 3.0 g/100g for black tea — awa-bancha and Ishizuchi-kurocha were lower than sencha and black tea.
Scope: Leaf analysis of awa-bancha and Ishizuchi-kurocha (both post-fermented teas), sencha (an unfermented tea), and black tea (a fermented tea)
Note: Included after checking the full text (Table 2) from the Tokushima Prefectural Industrial Technology Center. Details of sample size and origin/lot are not given.
Show data
ingredient-awa-bancha —has_composition→ "In an analysis conducted under the Food Labeling Standards (Cabinet Office Ordinance No. 10 of Heisei 27), tannin content was 8.1 g/100g for awa-bancha, 7.5 g/100g for Ishizuchi-kurocha, 11.4 g/100g for sencha, and 6.5 g/100g for black tea; caffeine content was 1.5 g/100g for awa-bancha, 1.4 g/100g for Ishizuchi-kurocha, 2.1 g/100g for sencha, and 3.0 g/100g for black tea. Awa-bancha had less protein, caffeine, and tannin than sencha, and less protein and caffeine than black tea."The same direction of change has also been observed in pu-erh tea, which undergoes pile fermentation. In a time series fermenting sun-dried green maocha from old-tree tea for 36 days, both bitterness and astringency were reported to be lower at the fermentation endpoint than in the raw material.
Scope: A 36-day fermentation time series of one sun-dried green maocha from old-tree tea
Note: A comparison between RM and F4. This is a time series from a single fermentation batch, not the causal effect or effect size of fermentation in general.
Show data
process-tea-pile-fermentation —was_observed_with→ "Lower bitterness and astringency than the raw material at the 36-day endpoint"Huigan, a time axis
Huigan (回甘, literally "returning sweetness") is the name of a sensation discussed in Chinese tea tasting. Tea-culture media articles repeatedly describe it as the sensation of feeling a slight bitterness and astringency immediately after taking a sip, after which sweetness becomes dominant over time, ending in a sweet aftertaste.
Scope: An explanation repeated across Chinese-language tea-culture media articles
Show data
concept-huigan —is_defined_as→ "In Chinese tea tasting, huigan is the sensation of feeling a slight bitterness and astringency immediately after taking a sip, after which sweetness gradually becomes dominant over time, ending in a sweet aftertaste."Several explanations have been proposed for why sweetness arrives with a delay. One commonly introduced hypothesis proposes a sequence in which salivary proteins and tea polyphenols first bind together in the mouth, and sweetness and moistness are felt when that binding comes undone.
Show data
concept-huigan —has_currently_proposed_mechanisms→ "The hypothesis (the "astringency-conversion" theory) that tea polyphenols bind salivary proteins to form a film in the mouth, perceived as an astringent, bitter-astringent sensation, and that when the film breaks, the oral muscles relax and huigan and shengjin (a moistening, salivating sensation) are felt."However, no study has yet been found that continuously measured the course from bitterness and astringency receding to sweetness rising. Most human studies measuring huigan are limited to rating its intensity at a single point after swallowing; no example measuring it as a continuous time curve, as with time-intensity (TI) methods or TDS, has been confirmed for tea. Even though claims can be gathered on where bitterness and astringency move across leaf, water, and fire, how sweetness rises afterward has not yet become a curve.
Note: Chong et al. (2019, 2022) both use a single-intensity rating method at a fixed timepoint (10 seconds after swallowing). There is an example of TI measurement for bitterness alone (theasinensin A), but it is not connected to sweetness or huigan. No study of huigan in tea using TDS was found within the search scope, and this is recorded here as an open gap in the evidence.
Show data
concept-huigan —has_methodological_limitation→ "Most human sensory studies of huigan are limited to a single-timepoint intensity rating, such as 10 seconds after swallowing; no example applying time-intensity (TI) methods or TDS to measure huigan as a continuous time curve has been found for tea."Frequently asked
Is green tea's bitterness the taste of caffeine, or of catechins?
As far as has been measured so far, reports centering on catechins as the main carrier predominate. In a study of green tea infusions, bitterness correlated strongly with the concentration of galloylated catechins. In an independent receptor experiment as well, the response to catechins bearing a galloyl group was far stronger than to those without one, and was reported to resemble the intensity order seen in human sensory studies. No study has been found that measured, using the same method as for catechins, how much caffeine itself contributes to green tea's bitterness. Conversely, there is a report that the umami compound theanine blocks one of the bitterness receptors, so bitterness is measured not as a single component but as a tug-of-war among several.
Is black tea's astringency caused by the same components as green tea's?
The results say no, not the same. One report finds that catechins and flavonol glycosides correlate strongly with astringency in green tea infusions, while a taste-dilution study on black tea identified a different group — neither catechins nor theaflavins — as the main contributor to astringency. Because the two studies differ in both the type of tea and the method, it cannot be separated out which one the discrepancy is due to. It has not been confirmed that the explanation "astringency is due to catechins" carries over as-is to black tea.
Why are matcha and gyokuro said to have a mild bitter-astringency?
The change in composition brought about by shading cultivation has been measured as the background. There is an account in which shading increases caffeine in the new shoots while decreasing catechins, with the two treated as separate items. Looking more closely, there is also a report that it is the weakly bitter-astringent epigallocatechin that decreases with shading, while the content of the strongly bitter-astringent epigallocatechin gallate barely decreases. In other words, as far as has been measured, the explanation that shading makes the tea milder is not because catechins overall decrease evenly, but because only the weaker one decreases selectively.
What is huigan, and how well confirmed is it?
Huigan is a word for the sensation of sweetness rising with a delay after the bitterness and astringency felt immediately on taking a sip, and it is repeatedly explained in the context of Chinese tea tasting. A hypothesized sequence has also been introduced in which salivary proteins and tea polyphenols bind in the mouth to form a film, and sweetness and moistness are felt when that film comes undone. However, no empirical study has been found that directly observed this film's formation and breaking itself. Most studies that measured huigan by human sensory testing are limited to a rating at a single point after swallowing, and no example measuring tea's huigan as a continuous time curve has been confirmed. It remains at the hypothesis stage.
What is not known
- No study has been found that directly isolated and measured, using the same method as for catechins, how much caffeine contributes to the bitterness of green tea and black tea respectively.
- Whether the main contributor to black tea's astringency works the same way in green tea's astringency has not been confirmed, since no study has directly compared across the teas studied.
- The mechanism by which huigan turns into sweetness still lacks support from direct observation in the mouth, and no measurement continuously tracking the passage of time has been found for tea either.
- Several reasons have been reported for why astringency falls in post-fermented tea, but no study directly linking fermentation conditions to the amount of decrease could be confirmed.
Read in the encyclopedia
The encyclopedia entries below are in Japanese.
- green tea infusion — the reference tea whose catechin and EGCG concentration is linked to bitterness
- tea catechins / epigallocatechin gallate — the compound group treated as the carrier of astringency and bitterness
- L-theanine — the umami compound that can antagonize a bitterness receptor
- tencha / matcha / gyokuro — teas whose catechin and amino acid composition is moved by shading cultivation
- Tieguanyin oolong tea infusion — an example where bitter-astringency moves with tea-to-water ratio, water temperature, steeping time, and number of infusions
- hojicha roasting / tea finishing fire (hi-ire) — processes in which the degree, temperature, and time of roasting remake catechins, caffeine, and aroma
- Goishi tea / Awa bancha / tea pile fermentation — teas whose catechins decrease with post-fermentation
- sweet aftertaste — huigan, the sensation, discussed on a time axis, of sweetness rising with a delay after bitterness and astringency
All entries in this hub (40)
Every encyclopedia entry that belongs to this hub, grouped by category. Entries without an English version open in Japanese.
ingredientAwa bancha · gyokuro · Kudingcha · Goishi tea · koicha (thick matcha) · ripe Pu-erh tea · Tieguanyin oolong tea infusion · tencha · matcha · yerba mate infusion · liangcha · green tea infusion
dishtereré
organismIlex kudingcha · yerba mate tree · Ligustrum robustum · Ilex latifolia
compoundL-theanine · epicatechin gallate · epigallocatechin · epigallocatechin gallate · caffeine
compound classtea catechins · flavan-3-ols · flavonol glycosides
sensationastringency
processmatcha milling · ishizuchi kurocha two-stage fermentation · tea shading cultivation · infusion time · tea pile fermentation · hojicha roasting · awa bancha boil-then-lactic fermentation · goishi-cha two-stage fermentation · tea finishing fire (hi-ire) · repeated sipping
concepthui tian (tea rising sweetness) · electronic tongue (taste sensing system)
Go deeper in the research notes
- Huigan — Tea's Time of Sweetness After Bitterness — follows the vocabulary of huigan, shengjin, and houyun, and the several hypotheses proposed to explain the sweetness
- Awa-bancha, Goishicha, Ishizuchi-kurocha — Japan's Post-Fermented Teas — compares the differences in process across three post-fermented teas, and the astringency and sourness a taste sensor shows against pu-erh
- Kuding Tea and Liangcha — Drinks That Talk About "Bitter" in Grades — covers bitter drinks called tea while coming from botanically unrelated plants
- Roasting Makes Bitterness, and Dark Roast Changes It — crosses over what the process of roasting does to the bitterness of coffee and hojicha
Source ledger
The primary information and limitations behind the claims listed here can be checked from the source link inside each card. The full list is gathered in the Source ledger.
How to cite this page
Aoki, M. (NIGAMI LAB) (2026). Tea's bitterness and astringency. https://nigamilab.com/en/topics/tea-bitterness/
Each claim card has a stable anchor (#clm-…). The dataset is CC BY 4.0; see the data page. Dataset DOI: 10.5281/zenodo.22739101