The Instrument That Detects Huigan Measures Friction
A tribology rig built to test machine lubricants finds that huigan, the returning sweetness after a good steep, tracks a measured friction coefficient on the tongue, at least for the tasters who can perceive it at all.
回甘, huigan, is the sweetness that arrives after the cup is down. Not the sweetness in the liquor. The one that comes back a few seconds later, at the root of the tongue and down the throat, and stays there. Gongfu practice is organized around it: the small vessel, the high leaf-to-water ratio, the short pours, all of it built to produce a finish rather than a mouthful.
Part of that sensation is not a taste. It is a mechanical event on the surface of the tongue, and it has been measured on the kind of rig engineers use to test lubricants.
Tribology is the study of sliding surfaces, and a mouth is one
Tribology measures friction, wear, and lubrication between two surfaces moving against each other. Engine bearings. Gear teeth. Hip implants. A tribometer presses one surface against another at a controlled load and speed and reports the friction coefficient, a dimensionless number describing how hard the pair resists sliding.
Point the same instrument at the mouth and the surfaces are the tongue and the palate, and the lubricant is saliva. Oral tribology is already an established way to study texture, because creaminess and fattiness are lubrication properties before they are flavors. Applying it to an aftertaste is the unusual move.
The correlation held below 0.5 mm/s, roughly the speed of a tongue
A group led by Pik Han Chong at Zhejiang Gongshang University published the result in Food Research International in November 2019: "Oral" tribology study on saliva-tea compound mixtures1. Twenty-four people first rated huigan intensity across a set of tea compounds. Twelve trained panelists were then carried forward and split six and six by sensitivity.
Among the sensitive panelists, rated huigan intensity tracked the measured friction coefficient, and the correlation was strongest at sliding speeds below 0.5 mm/s. That speed is the point of the finding. Slow sliding is the boundary regime, where the lubricating film thins and the two surfaces come close to touching, and it is roughly the speed at which a tongue works against a palate.
The compounds tested were epicatechin (EC), epigallocatechin (EGC), L-theanine, and trilobatin.
The compound that produced the strongest huigan also produced the highest friction
The follow-up, Tea compound-saliva interactions and their correlations with sweet aftertaste2, appeared in npj Science of Food on 9 February 2022. An EC/EGC mixture at 400 and 100 ppm respectively rated 5.2 out of 11 for huigan without swallowing, against 2.8 for plain water. With the swallow, 5.2 against 3.1.
The same mixture raised total salivary protein secretion significantly in both groups. Salivary lipase activity fell after most samples, significantly so for EC/EGC in the sensitive group.
And the same mixture showed a high friction coefficient at low sliding speed. High oral friction is what astringency is: catechins bind salivary proteins, the proteins stop lubricating, and the mouth dries and grips. The two studies do not collapse astringency and huigan into one sensation, but they do trace both to the same protein-polyphenol interaction in the mouth. The tea that grips is, in this account, the tea that returns sweet afterward.
Half the trained panel did not perceive it at all
Six of the twelve panelists were classified as sensitive, meaning they perceived huigan correctly across tea brews and compound solutions. The other six did not. The friction correlation lives in the sensitive group.
So the defensible statement is narrower than the headline. It is not "huigan is friction." It is: among people who perceive huigan reliably, the intensity they report tracks a mechanical measurement of their own saliva. Whether the non-sensitive group lacks the sensation or lacks the vocabulary for it, the studies do not resolve.
A novice who finds nothing returning in the throat is not necessarily brewing badly.
None of this has been measured on oolong
Say the limit plainly. EC and EGC are catechins studied here at concentrations chosen for a green-tea-style profile, and the sensory work was done on compound solutions, not on a partially oxidized leaf.
Partial oxidation is the whole definition of oolong. Polyphenol oxidase converts a fraction of the catechin pool into theaflavins, thearubigins, and related dimers, so an oolong is poured with a different polyphenol distribution than a green tea carries. The mechanism is a protein-binding, surface-friction argument, and oxidized polyphenols bind salivary proteins too, so it plausibly carries over, but it has not been demonstrated in oolong specifically in this literature.
It is also one contributing mechanism, not the explanation. The chemistry of roasting shifts an oolong's aftertaste by a separate route, covered here already. This piece concerns what happens apart from that: the physical feel of the sip itself.
A 2026 review in Comprehensive Reviews in Food Science and Food Safety, Sweet Aftertaste (Huigan) in Foods3, states plainly that the mechanisms behind the sensation and its material basis remain inadequately elucidated. Nobody has closed the file.
Chinese tea drinkers described it as a bodily event first
Popular Chinese writing on huigan has long explained it physiologically rather than as a flavor note: organic acids provoking salivation, tea polyphenols forming a film across the oral mucosa that then breaks, polysaccharides meeting salivary amylase. That is folk physiology, circulated in tea-culture writing, not peer-reviewed evidence.
But the shape of it is worth marking. It treats huigan as something happening to the mouth, not something arriving at a taste receptor. Western laboratory work has now put an instrument behind that shape.
What this changes at the table
If perceived huigan tracks the friction of a saliva-polyphenol mixture, then the levers are the polyphenol load you deliver and the state of the saliva meeting it.
Do not chase the returning sweetness with a thin infusion. A weak steep delivers little to bind, and there is nothing to grip and release.
Judge the finish with an empty mouth. No water, no palate cleanser, thirty seconds at minimum. The measurement window opens after the swallow, and rinsing between steeps resets the very salivary protein state the effect depends on.
And by the seventh steep the extractable polyphenol load has fallen sharply. A huigan still returning at that point is reporting on the leaf, not on your brewing.