The Science of Hui Gan: Where Tea's Returning Sweetness Really Comes From

Tea Science｜Author：祥興茶行｜Published：2026-07-13

The sweetness that rises in the throat after a bitter, astringent sip comes from overlapping events: astringency, a tactile sensation, arrives first and fades, while the sweet compounds the tea carried all along get noticed. Here is what the research supports and what is still analogy.

Take a sip of strongly brewed tea and the surface of your tongue tightens first, with a bitter, drying edge. Fifteen or twenty seconds later, a clean sweetness slowly rises from the throat and the back of the tongue. Many people fall for tea because of that arc: wince first, unwind after. This article looks at whether that sweetness, the hui gan (回甘) tea drinkers prize, was in the liquor all along, or is thrown into relief once the bitterness and astringency recede.

Hui gan can be broken into several things happening in the mouth at the same time: a layer of astringency coats the tongue first and then slowly dissolves, while the tea's own sweet compounds are present throughout and only now get tasted. The two are staggered in time, and that gives you "bitter and astringent first, sweet after."

The thread of this article: hui gan, split into three events ① Astringency arrives first (a tactile sensation) → ② the astringency slowly fades → ③ the sweetness and umami that were in the liquor all along now stand out. Hui gan is these three events layered together in the mouth.

Astringency isn't actually a taste

Astringency is touch, not taste.

Taiwan's Tea and Beverage Research Station, under the Ministry of Agriculture, explains it this way: the tannins in tea (polyphenols) have a strong affinity for the proline-rich proteins in saliva. When the two meet, they bind and precipitate, laying a thin film across the inside of the mouth. Nerve endings sense the pressure and friction of that film and send the signal to the brain, and we "taste" astringency [1]. So astringency is the feeling of the tongue being gently cinched, unlike chemical tastes such as sweet, sour, bitter, or salty. The main source is the catechins, which come in esterified and free forms, both bitter and astringent [1].

〔圖〕Diagram of tea polyphenols binding salivary proteins to form a film

One research team coated "artificial oil bodies" with recombinant protein to stand in for salivary proteins, then measured how strongly various tea polyphenols bound to them. The ester-type catechins, which carry a galloyl group, were more astringent than the free forms (the rough ranking was ECG > EGCG > GCG > CG), and as steeping time lengthened, astringency rose exponentially rather than in gentle proportion [3].

An in vitro model: the "artificial oil body" setup is an in vitro model that uses recombinant proteins to simulate saliva; it was not measured in a human mouth. It shows the direction of the mechanism and relative strengths, but it is not in vivo evidence [3].

Catechins are not the only source of astringency. A taste study of black tea used taste dilution analysis to compare compounds one by one, and found that the main drivers of black tea's astringency are neither the large thearubigins nor the catechins or theaflavins, but a group of trace flavonol glycosides, the most important being rutin [4]. Its measured astringency threshold was as low as 0.001 µmol/L, 190,000 times lower than EGCG, the main bitter-astringent compound, and 16,000 times lower than the theaflavins [4]. A minute amount is enough to leave a velvety grip across the mouth.

So astringency is not just a matter of how much polyphenol there is. The shape of the molecule and which sugar is attached matter more than quantity (the same quercetin carrying different sugars can differ in astringency threshold by a factor of several hundred) [4]. This is black tea research; green and oolong teas have different compositional profiles, so it cannot be applied directly. It also measured only astringency thresholds, not sweetness or hui gan, so it helps explain astringency but is not evidence about hui gan [4].

The sweetness was in the cup all along

The most concrete source of hui gan is that tea liquor really does contain sweet compounds. They are often masked by the early bitterness and astringency and get noticed half a beat late.

Ministry of Agriculture materials divide the taste of tea into four types: sweetness comes mainly from sugars, with some amino acids contributing; umami comes mainly from theanine and glutamate-type amino acids; bitterness comes mainly from caffeine (catechins are bitter too); and astringency comes mainly from catechins [2].

Taste ｜ Main compounds

Sweet ｜ Mainly sugars; some amino acids contribute [2]

Umami ｜ Theanine, glutamic acid, other amino acids [2]

Bitter ｜ Mainly caffeine; catechins carry bitterness too [2]

Astringent ｜ Catechins (esterified and free) [1][2]

The official FAQ gives two frequently cited ballpark shares for the main compounds:

On the link between bitterness, astringency and sweetness, one green tea experiment is useful. Researchers used tannase to hydrolyze the esterified catechins in the infusion (EGCG and ECG, the main bitter-astringent compounds), and both the "sweet aftertaste" and the overall acceptability of the liquor rose [6]. This fits the inference that bitter-astringent intensity suppresses our perception of sweetness. What it shows is that removing bitterness and astringency makes the sweet aftertaste more apparent; it does not mean the breakdown products of catechins are themselves sweet.

For pu-erh, a 2025 study of raw pu-erh (Yuan et al., published in Food Chemistry) combined sensory evaluation, mass-spectrometry quantification and spiking experiments to track which compounds drive hui gan [7]. The two key compounds most strongly correlated with hui gan intensity were theophylline (an alkaloid) and rutin (a flavonol glycoside); neither is a sugar [7]. When the researchers added theophylline to infusions with weak hui gan, hui gan clearly strengthened within a range of roughly 182.9–290.4 micrograms per 250 mL; beyond the upper bound, sweetness was suppressed and bitterness and astringency came through, a classic too-much-of-a-good-thing dose response [7]. Their stepwise regression models had high explanatory power (R² roughly 0.91–0.99) [7].

What this study covers: it measured correlations between compound levels and hui gan intensity, plus causation of the form "add this compound and hui gan strengthens." It did not directly observe astringency fading while sweetness rises in step. It shows that certain compounds affect hui gan, but it is not enough to support a full mechanism. How these compounds produce sweetness, for example the speculation that salivary enzymes break down rutin and the products activate the sweet receptor TAS1R2/TAS1R3, or that theophylline raises sweetness indirectly through taste contrast and increased salivary flow, remains inference from the literature, not something this study measured [7].

The timeline: bitter first, sweet after

Put those two pieces on one timeline:

〔圖〕Timeline schematic of hui gan: astringency rises then falls while sweetness surfaces late

A schematic: it shows the direction of the trends, not measured curves. The red line is astringency, rising quickly after the sip and then slowly dissolving. The green line is the perception of sweetness and umami, slow to start and distinct only later. The late stretch where the two lines cross is when hui gan is felt.

Astringency fades because of saliva. A physiological model of tea-induced astringency holds that astringency builds up sip after sip, and that it takes time to subside, along with salivary proteins being replenished and the mouth being re-lubricated [5]. This matches the research station's description of the film: saliva keeps flowing and gradually carries the complexes away, the film dissolves, the tight feeling releases, and the sweetness from the sugars and amino acids that were there all along shows through. The model only establishes that astringency needs time to recover; there is not yet complete data on how long or how fast it fades, so no specific number of seconds is given here.

The "bitterness fades, sweetness by contrast" explanation

Another explanation holds that the tea does not turn sweet at all. The mouth adapts to the bitterness and astringency, and once the stimulus is removed, even plain water tastes slightly sweet; hui gan is a contrast effect.

This idea has a basis in taste science. Classic psychophysics experiments show that human taste adapts: after sustained exposure to a taste, plain water itself produces an aftertaste. In salt-adaptation experiments, for example, solutions of different concentrations taste shifted across the board after adaptation [8].

These experiments did not use tea: the adaptation and contrast experiments above used table salt and other standard taste-research materials. Applying them to hui gan is a mechanistic analogy and a reasonable inference, not evidence from tea itself [8].

So the two explanations do not rest on equal evidence:

Two explanations, two grades of evidence ・"The liquor carries its own sweetness, tasted half a beat late": supported by first-hand tea studies and official data (the green tea enzymatic-hydrolysis experiment [6], the official compound-to-taste mapping [2]). However, the latest raw pu-erh study [7] found that the key hui gan compounds are theophylline and rutin, not sugars, so this camp has not settled on "the sweetness comes from sugars" either. ・"Sweetness thrown into relief as bitterness fades": so far, no study has directly measured, in tea, sweetness perception rising in step as astringency decays. It relies on general taste-adaptation principles [8] plus the model that astringency fades over time [5]. It is a reasonable inference that still lacks a direct, controlled measurement in tea.

A practical way to put it: hui gan is likely a combination of compounds that are sweet in themselves and the contrast of bitterness and astringency fading. The first is well supported; the second is still an analogy, and the two do not carry equal weight of evidence.

Hui gan is more than "sweet": there's body, too

When people describe hui gan, they use words beyond "sweet," such as thick, smooth, and a long finish in the throat. In taste science this corresponds to kokumi (a sense of fullness and continuity in the mouth). Research has identified kokumi-active taste peptides in green tea (γ-Glu-Gln) and glutathione (GSH), and when glutathione and theanine are present together, the sense of body strengthens markedly [10].

Kokumi and sweetness: kokumi accounts for the lingering, full, thick dimension of hui gan, not for the source of the sweet taste itself [10].

Common claims about hui gan

Here is where the evidence stands on two claims you often hear.

Claim one: "If it has hui gan, it's a good tea" There is some compositional basis: teas with lower bitterness and astringency, where sugars and amino acids are easier to taste, do make hui gan easier to notice [9]. But the equation "hui gan = good quality" currently has no direct support from primary research. Hui gan is a useful reference when tasting, not a sole measure of quality.

Claim two: "Hui gan means sugar was added, or pesticide was sprayed" Claims like these circulate, but there is neither primary scientific evidence supporting them nor evidence refuting them. With no evidence either way, it is too early to conclude in either direction.

The takeaway, in one cup

References

  1. 陳國任（撰稿）、楊美珠（諮詢）。〈茶之澀味〉。農業部茶及飲料作物改良場。https://www.tbrs.gov.tw/ws.php?id=4825
  2. 農業部。〈茶湯滋味與主要成分〉常見問答（faq id=220、221）。農業知識入口網。https://www.moa.gov.tw/faq_view.php?id=221 、 https://www.moa.gov.tw/faq_view.php?id=220
  3. Liu C-T, Tzen JTC (2022). Exploring the Relative Astringency of Tea Catechins and Distinct Astringent Sensation of Catechins and Flavonol Glycosides via an In Vitro Assay Composed of Artificial Oil Bodies. Molecules 27(17):5679. DOI 10.3390/molecules27175679
  4. Scharbert S, Holzmann N, Hofmann T (2004). Identification of the Astringent Taste Compounds in Black Tea Infusions by Combining Instrumental Analysis and Human Bioresponse. Journal of Agricultural and Food Chemistry 52(11):3498–3508. DOI 10.1021/jf049802u
  5. Nayak A, Carpenter GH (2008). A physiological model of tea-induced astringency. Physiology & Behavior 95(3):290–294. DOI 10.1016/j.physbeh.2008.05.023
  6. Zhang Y-N, Yin J-F, Chen J-X, et al. (2016). Improving the sweet aftertaste of green tea infusion with tannase. Food Chemistry 200:150–157. DOI 10.1016/j.foodchem.2015.07.046
  7. Yuan X, Chen X, Chai C, Feng M, Hu Y, Yi Z, Gu Y, Ruan L, Yi L (2025). Identifying key contributors to the sweet aftertaste of raw Pu-erh tea through analytical and sensory methods. Food Chemistry 481:144067. DOI 10.1016/j.foodchem.2025.144067
  8. Bartoshuk LM, McBurney DH, Pfaffmann C (1964). Taste of Sodium Chloride Solutions after Adaptation to Sodium Chloride: Implications for the "Water Taste". Science 143(3609):967–968. DOI 10.1126/science.143.3609.967
  9. Zhang L, Cao Q-Q, Granato D, Xu Y-Q, Ho C-T, et al. (2020). Association between chemistry and taste of tea: A review. Trends in Food Science & Technology 101:139–149. DOI 10.1016/j.tifs.2020.05.015
  10. Lu J, Cao Y, Pan Y, et al. (2022). Sensory-Guided Identification and Characterization of Kokumi-Tasting Compounds in Green Tea. Molecules 27(17):5677. DOI 10.3390/molecules27175677

Teas mentioned