日本語版

Research note 054 — 2026-09-07

The Science of Aku Removal Has Not Measured Bitterness

In the kitchen, saying 「アクを抜く」 (in translation: "remove the aku") bundles together changes in bitterness, egumi, astringency, and color all at once. Boiling sansai in water with ash or baking soda, simmering bamboo shoots with rice bran, blanching and discarding the water for green vegetables — all of these line up under the single phrase 「アク抜き」 (in translation: "aku removal"). Convenient as it is as a prep step, this single word conceals what is actually moving. But food sensory-evaluation materials do not measure this single word as such. Egumi and bitterness have, from the outset, been set up in several studies as separate evaluation items.

What this page establishes

• Bitterness and egumi have, from the outset, been measured as separate evaluation items in multiple studies.
• The 1980 review had already concluded that research on how bitter compounds change through cooking has hardly been carried out.
• How much glucosinolate leaches into the boiling water when vegetables are boiled has reported recovery rates that diverge widely between studies, and remains unsettled.
• Whether sansai aku removal (pretreatment using ash, rice bran, or baking soda) moves bitterness itself has not yet been measured.

Egumi and bitterness are measured as separate things

In a dictionary of evaluation terms for cooked rice, egumi is defined as a husky sensation as if aku were clinging to the tongue and throat, and is placed as an item separate from bitterness.

In an evaluation of soy sauce, by contrast, astringency and egumi are bundled into one item, while bitterness and its aftertaste are separated out as their own item. This is a case where two different sensations, astringency and egumi, coexist inside a single word. Astringency is tied to the tongue-puckering sensation from eating an astringent persimmon or to the distinctive taste of hard water and nigari, bitterness to the taste of caffeine, and bitterness aftertaste is redefined, in its own item, as the bitterness still remaining on the tongue 10 seconds after tasting the soy sauce.

In spinach, when the water-extracted components were passed through an adsorbent, they separated into two groups with different properties, and bitterness concentrated in one of them.

Building on this separation, the conclusion drawn was that spinach's bitterness is carried by a compound other than oxalic acid, which has been named as a candidate for egumi. What has been spoken of as the cause of egumi and bitterness itself are not the same substance.

What the 1980 review had already said

Looking back with this separation in mind, the stance of a review of bitter compounds written in 1980 comes into view. This review tabulated the human thresholds of 10 bitter substances lined up from urea to brucine, comparing the bitterness intensity of each compound.

Yet the same review's closing words recorded that research on the formation of bitter compounds in plant-based foods, and on how they change through cooking, has hardly been carried out. Even though the bitterness intensity of a single compound could be tabulated, how the act of cooking moves that bitterness had already, by 1980, been named as a gap.

What has been measured since is leaching, not removal

What advanced after the review was not the bitterness score itself, but research on the mechanism of what leaches into the water during boiling. In an experiment boiling Brassica vegetables, one report found that roughly 90% of the lost glucosinolates were recovered from the boiling water, while in a different experiment only 7% of the aliphatic types and 18-20% of the indolyl types were recovered from the boiling water, and even combined with the heated sample this did not reach 100%. A portion remains unexplained by leaching alone, and the mechanism itself does not agree between studies. Because vegetable type, cutting, and the temperature reached during heating differ between studies, the two recovery rates cannot simply be compared.

In lupin seeds too, soaking and boiling remove 89-97% of the bitter alkaloids from the seed, and a process as short as 1 hour has been reported. Traditionally, a process of soaking in salted water and washing is known, and alongside this, multiple methods — biological processing, chemical processing, and water-only processing — have been reported in parallel. What is measured here is the amount of chemical compound remaining in the seed, not the bitterness score itself as tasted in the mouth.

Fukinoto's bitterness resists water

For fukinoto, there is a report that almost no bitterness transfers into either a hot-water extract or a dilute-alcohol extract. In response to this result, the researcher who reported it left behind the view of bitterness as a glycoside and switched to extraction with an organic solvent.

What this switched extraction yielded were isopetasin, with a melting point of 95°C and accounting for 0.09% of the air-dried plant, and fukinone, at a 0.2% yield — an oily substance and a crystalline substance, respectively, both described as bearing bitterness. Given that substances with poor water solubility carry at least part of fukinoto's bitterness, how far exposing it to hot water moves that bitterness cannot be read off from the extraction results.

Research on masking has advanced

What has been measured as an operation that actually moves bitterness itself is closer to masking than to removing. In an experiment adding sugar at 4 levels from 0-1.8%, or salt at 0-0.2%, to purees of broccoli, spinach, and kale, bitterness intensity dropped in all 3 vegetables, while other sensory attributes such as texture and aroma did not change.

On the other hand, in an experiment soaking pieces of bitter gourd (goya) fruit in 3.5% salt water at 25°C for 60 minutes, the bitterness score barely changed from the untreated fruit. However, the treated batch also went through a hot-water blanch before squeezing, so the effect of the salt water itself is not fully separated out. Even under the same idea of soaking in salt, how well it works on bitterness diverges depending on the subject and the conditions.

Sansai prep remains an unverified row

Blanching-and-discarding the water for fukinoto, rice bran for bamboo shoots, ash or baking soda for warabi (bracken). In the prep ledger, these line up as 「未検証」 (in translation: "unverified") rows with an empty verdict column. The column is empty not because no source was searched for, but because one was searched for and not found, and the story behind each row is linked, row by row, to an entry in the hypothesis box. The shape of the row is fixed: ingredient, method, condition, targeted compound, verdict, source. The moment a claim enters the source column, that row moves from unverified to one of「減る」 (in translation: "decreases"), 「変わらない」 (in translation: "unchanged"), or 「増える」 (in translation: "increases").

The experiment needed to make that move is not large-scale. Buy the same ingredient on the same day, and change just one variable — the number of blanch-and-discard cycles, whether rice bran is used, the amount of baking soda — to make two or three conditions. Score bitterness intensity on the same scale, and keep a record of the date and the ingredient's provenance. This pattern is the same as 「家で確かめる」 (in translation: "check it at home") on the bitter crossings, and the record goes into the measurement-ledger table.

The gap can be filled in the kitchen

The gap pointed out by the 1980 review was not only how bitter compounds change through cooking. In the same closing words, it was written that research on the formation side — how bitter compounds in plant-based foods arise in the first place, and how to prevent that — also had much left to wait for. Cooking-side research has since filled part of this gap, and how far the formation-side gap has been filled needs to be checked separately.

On the cooking side too, this gap can be filled in the kitchen. Prepare the same ingredient, change only one prep condition, taste-compare, and write down bitterness intensity each time, in words or numbers. Moving the number or duration of blanch-and-discard cycles, the length of soaking in water, or whether rice bran is used, one at a time, and recording it, becomes the first step toward filling in a row this ledger does not yet have.

Frequently asked

Does aku removal reduce bitterness?

Whether traditional aku removal using ash, baking soda, or rice bran moves the amount of bitterness itself has not been confirmed within the scope of this material. What has mainly been measured is a different operation: in an experiment soaking bitter gourd in salt water, the bitterness score barely changed from the untreated fruit, while in an experiment adding sugar or salt to purees of green vegetables such as spinach, the reported result was a drop in bitterness. These operations are closer to 「隠す・和らげる」 (in translation: "masking, softening") than to 「抜く」 (in translation: "removing") bitterness, and a measurement of bitterness targeting the aku-removal procedure itself has not yet been found.

How do aku (egumi) and bitterness differ?

In food sensory evaluation, egumi and bitterness are often set up as separate evaluation items. A dictionary of evaluation terms for cooked rice defines egumi as something felt as a husky sensation remaining on the tongue and throat, and sets it up as an item separate from bitterness. In spinach, when components were separated into those dissolving into water and those dissolving into ethanol, bitterness concentrated in the latter, and it has been reported as a compound other than oxalic acid, which has been named as the cause of egumi. However, this separation differs by subject and by study — in soy sauce evaluation there is also a case in which astringency and egumi are bundled into one item — so there is no single, uniform definition.

What is not known

• Whether aku removal using ash, baking soda, or rice bran, respectively, actually reduces the amount of bitterness was not found within the scope of this material.
• Whether the recovery rate of glucosinolate into the boiling water can be explained by leaching alone, and how much thermal degradation is involved, remains unsettled, with results diverging between studies.
• An experiment directly measuring how much fukinoto's bitter substances (isopetasin, fukinone) move under exposure to hot water was not found.
• What has been measured for lupin debittering is the amount of compound remaining in the seed, and no material has been reached that also measured the bitterness score as tasted in the mouth through the same process.
• For sansai other than spinach and bitter gourd, an experiment comparing bitterness while changing prep conditions one at a time was not found within the scope of this material.

Related: egumi (Japanese food descriptor) /fukinoto (butterbur flower bud) /aku-nuki (blanch-and-soak treatment) /the prep ledger /the hypothesis box /Research note 053: Rereading Kitchen Prep Through Measured Bitterness