日本語版

Topic hub — 2026-09-07

Citrus bitterness

Citrus bitterness is not a single, monolithic sensation decided the instant the fruit is squeezed. The compounds that keep coming up as its carriers are naringin and limonin: the former is linked, in many citrus products, to an immediate bitterness present from the moment of pressing, and the latter to a delayed bitterness that arises some time after pressing. The delay happens because non-bitter precursors slowly convert into bitter compounds inside the juice, and what drives that conversion has not been settled between sources. Bitterness is also not spread evenly through the flesh; measurements find naringin concentrated in the white pith and membranes, and limonoids in the outer peel and seeds. Several ways of removing it are in use — enzymes, resins, salt water, high pressure — and none of them has been shown to erase only the bitterness; resins also adsorb some aroma compounds (though the one reduced most is described as an off-flavour). The differences in content between cultivars carry straight through into how they get talked about as products. Here, each of these is laid out separately, without merging them.

What this page establishes

• That citrus bitterness has two kinds — one that stands up immediately and one that arrives late — has been confirmed both by reviews and by individual kinetic studies.
• The phenomenon of juice growing more bitter after pressing has been confirmed by multiple studies as a chemical conversion in which a non-bitter precursor turns into limonin after pressing. What drives the conversion, however, has only been studied with different materials and measurements; no study has yet compared acid and enzyme under the same conditions inside the juice.
• Several measurements that compared parts of the fruit by the same method confirm that naringin is concentrated in the white pith (albedo) and the membranes. Limonoids are measured as higher in the outer peel, membranes and seeds than in the white pith, so where bitterness gathers differs by compound.
• In experiments using resin, some aroma compounds were adsorbed along with the bitter compounds. The one reduced most is described as an off-flavour, and whether the overall aroma got worse was not measured. A way to remove only bitterness selectively has not yet been measured.

The carriers of bitterness

In experiments that rebuilt orange juice bitterness from its components, no single indicator emerged that alone would reveal the bitterness. The result was that bitterness is decided by a combination of limonoids and their glycosides, polymethoxyflavones, organic acids, and sugars.

Within that interplay, the names that keep coming up are naringin and limonin. Naringin is positioned as a flavanone glycoside that contributes to immediate bitterness in many of the citrus products where it occurs.

Limonin, by contrast, is a delayed-bitterness limonoid: at the moment of pressing it exists in the form of a non-bitter precursor (limonoate A-ring lactone), and it appears only after converting, over time, into limonin itself. The same juice holds, side by side, something that is already bitter and something that is about to become bitter.

The bitterness that arrives late

The term "delayed bitterness" is not a figure of speech; it names a chemical conversion. In citrus juice after pressing, the process by which the non-bitter precursor limonoate A-ring lactone (LARL) converts into bitter limonin is, itself, what the delayed bitterness consists of.

Sources differ on what drives this conversion. A kinetic study that tracked seedless navel orange juice takes it to be an acid-promoted reaction, while a study that purified an enzyme from citrus seeds frames it as catalysis by limonin D-ring lactone hydrolase (LDLH). The two differ in material (seedless juice versus seeds) and in how they measured, so the findings are not in conflict so much as not yet directly compared. No source has confirmed which is the main driver inside the juice itself.

The concentration of limonin itself also moves with time after harvest. In a study that tracked navel orange juice, the post-pressing maximum value of limonin fell the longer the fruit had been stored after harvest, dropping from about 24.24 to about 1.04 mg/L over roughly 0 to 45 days. These values come from fruit of one harvest kept at 20 to 25 °C and 70 to 90% relative humidity, with the pressed juice given 0.1% potassium sorbate and held at 25 °C, so they do not transfer directly to home storage. Heat processing and acidification only sped up the formation; they did not change the maximum value itself.

The concentration at which it can be perceived as bitter is also low. Limonin's bitterness recognition threshold in distilled water is 1 mg/L, and in orange juice at pH 3.8 the maximum recognition threshold is 6.5 mg/kg.

Where it concentrates

Looking at the origin of bitterness not just by compound but by location in the fruit, the white pith stands out. In grapefruit, the literature describes naringin as being extracted into the juice from the membranes and the albedo (the white pith). Japanese measurements that compared parts by the same method also found the naringin of natsumikan and hassaku concentrated in the white pith and membranes and low in the flesh and juice, while limonoids were higher in the outer peel and seeds than in the white pith.

A study on the white mesocarp (albedo) of pomelo quantified the naringin and limonin content itself before and after processing. With combined β-cyclodextrin and ultrasound processing, naringin fell from 18.90 mg to 5.1395 mg per gram, and limonin from 0.2907 mg to 0.0776 mg.

In the same study's sensory evaluation, the minimum post-processing bitterness score was 1.31 (the number and training of the assessors are not stated in the methods; the 15 people were participants in a separate post-meal test). The origin of that number remains at a preliminary standing. That processing the white pith lowered its bitter compounds is recorded by quantification, but the sensory side rests on 1 study with an inconsistent scale and an unknown number of assessors.

Ways to remove it

One way to remove it is a process using enzymes. A process combining immobilized naringinase with the simultaneous adsorption of limonin onto agarose modified with butyl or octyl groups reduced naringin by 54-74% and limonin by 76-100%, under conditions of pH 4.0 and 30°C.

Adsorption by resin is also used. Static adsorption tests selected the macroporous resins PAD550 and PAD600, and fixed-bed processing with PAD550 lowered naringin below the taste threshold and also reduced limonin, nomilin, and bergamottin. In an experiment using a different resin, XAD-7HP, naringin fell to 51.28% of its initial value, and the aroma compound alpha-terpineol fell to 43.05% (the other aroma compounds kept 76.27 to 92.76%). The authors describe alpha-terpineol as an off-flavour and see its selective removal during debittering as promising. The overall quality of the aroma was not measured in this experiment.

Processing with salt water also moves bitterness. In jam made with bitter-orange peel, the batch processed with water alone matched the bitter control, and only the batch processed with salt water lowered its bitterness rank. This compares the whole process of repeated soaking and heating, though, not the effect of salt water alone. In terms of compounds, most of the naringin in natsumikan peel dissolved out after 2 to 3 hours of soaking in either salt water or plain water, and boiling also cut it sharply.

There is also the option of making it without heat. In an evaluation by 27 students and staff who compared the same formulation of yuzu-peel marmalade made by a high-pressure method and by a heating method, the high-pressure method scored lower on bitterness, but what was shown as significant was the liking of the bitterness, the texture and the overall evaluation; the difference in bitterness strength is not stated as significant. Neither method included a debittering step such as soaking.

How Japan's mid-to-late citrus varieties get talked about

Every one of the mid-to-late citrus varieties whose defining trait is bitterness is talked about, separately from any record of its compounds, together with a story of discovery or origin. Hassaku carries an account of being discovered at a temple on Innoshima in Hiroshima City, and a naming story in which a monk said it could be eaten around the first day of the eighth month of the lunar calendar; but the source notes this is out of step with when the fruit is actually at its best, and that the true origin of the name itself is unclear.

Kawachi bankan is described as a chance seedling of the pomelo lineage discovered in Kawachi village, Kumamoto Prefecture. Official materials record an account of discovery in 1935 — other sources give the year as 1934 — followed by its arrival in Ainan, Ehime Prefecture in 1971, where a warm climate suited long, on-tree ripening.

Natsumikan is talked about, in Hagi City's official materials, tied to the town's finances. The account is that after being brought to Hagi as a substitute for yuzu, cultivation spread through Meiji-era policy, and by Meiji 22 revenue from natsumikan had come to exceed the scale of the town's own finances.

Amanatsu is talked about as a branch sport of that natsumikan. It is described as having been found as a branch sport (bud mutation) in the early Showa era, at a summer-citrus orchard in Tsukumi City, Oita Prefecture, with the added feature that its citric acid drops faster than ordinary natsudaidai, shifting when it is at its best.

How beverages describe themselves

These differences in how varieties get talked about carry straight through into how processed products describe themselves. As one example of a mass-market product, the official product descriptions for Morinaga Milk Industry's mainstream Sunkist 100% citrus juices do not use the word "bitterness" at all. Yet within the same brand, "Rich Grapefruit" explicitly states a hint of bitterness — the wording splits even inside one brand ("Rich Grapefruit" comes from the announcement of a limited-time product sold in 2015, not from the same period as the current regular products).

Suntory Natchan Orange's official product information explains that it achieves both the fruit's natural juiciness and easy drinkability with sourness and bitterness held down, positioning bitterness as something to be suppressed.

The wording changes, though, once you reach growers selling direct from the producing region. Itonoen describes kawachi bankan as having "a refreshing hint of bitterness" and hassaku as having "a hint of bitterness that's addictive," explicitly presenting bitterness, for some cultivars, as an active selling point.

There is also an example, among processed lemon products, of bitterness talked about as a value. Pokka Sapporo's official Kiretomon Lemon brand site explains its use of lemon peel extract by saying it "also adds the bitterness of the lemon peel, so we deliver the value of the lemon to our customers without leaving anything out," positioning bitterness as part of the lemon's value.

Frequently asked

Does orange juice get more bitter over time after it's squeezed?

The phenomenon of limonin concentration rising after pressing has been confirmed in multiple studies. It has been reported that in freshly squeezed navel orange juice, limonin content rises during storage and reaches a fixed maximum value. This is explained as happening because what existed, before pressing, in the form of a non-bitter precursor converts, after pressing, into the bitter substance limonin. What mainly drives that conversion, however, is not agreed upon. A kinetic study using seedless fruit takes it to be an acid-promoted reaction, while a separate study that purified an enzyme from seeds takes it to be enzymatic catalysis. Neither source has directly confirmed, inside actual juice, which of the two is the main driver.

Why are the white strings in mandarins and grapefruit bitter?

Whether bitter compounds gather in the white pith depends on the compound. Measurements that compared parts of the fruit by the same method found the naringin of natsumikan and hassaku concentrated in the white pith and membranes and low in the flesh and juice, while limonoids were higher in the outer peel, membranes and seeds than in the white pith. In grapefruit, the literature describes naringin as being extracted into the juice from the membranes and the white pith. No study was found that compared, by taste, the bitterness of the white pith alone with other parts.

Can citrus bitterness be removed by soaking, salt water or high pressure?

How well each method works has been measured. In terms of compounds, the naringin of natsumikan peel was measured to drop sharply with soaking (in salt water or plain water) and with boiling. In sensory terms, in jam made with bitter-orange peel the water-only processing matched the bitter control, while only the batch processed with salt water lowered its bitterness rank (a comparison of the whole process of repeated soaking and heating). In an experiment comparing yuzu-peel marmalade made by a high-pressure method and a heating method, the liking of the bitterness and the overall evaluation were significantly higher for the high-pressure method, but the difference in bitterness strength is not shown as significant. Industrial methods using enzymes or resins have been confirmed to reduce bitterness substantially in juice. For peel, though, a study found that banpeiyu peel treated with naringinase was rated more bitter than peel treated with hydrochloric acid or hot water. Resins also adsorbed some aroma compounds, but the one reduced most is described as an off-flavour. The conditions and results of each study, by preparation step, are laid out in the comparison Does boiling or soaking cut citrus peel bitterness?

Why are hassaku and kawachi bankan described as having bitterness as their defining trait?

Measurements that compared bitter-compound content across cultivars by the same method do exist: the naringin in the juice was highest in natsumikan, then Kawano natsudaidai, then hassaku. No measurement has been reached for kawachi bankan. Looking at how they are described, what is known is that these mid-to-late citrus varieties have been talked about together with stories of discovery and origin. Hassaku carries an account of being discovered at a temple on Innoshima in Hiroshima City, but the source notes that the timing carried in the naming story is out of step with when the fruit is actually at its best, and that the truth of the origin itself is unclear. Kawachi bankan is talked about, in official materials, as a chance seedling of the pomelo lineage discovered in Kawachi village, Kumamoto Prefecture, and natsumikan as a crop that supported Hagi's finances. The language that describes bitterness as a defining trait comes not from measuring compounds but from the side of local origin stories and how the products are talked about.

What is not known

• Whether acid or an enzyme is the main driver of limonin formation after pressing has not yet been measured — no study has been found that confirmed this directly inside the juice itself.
• No study has been found that compared, by taste and under the same conditions, the bitterness of the white pith with other parts of the fruit (compound content has been compared across parts by the same method).
• No study has been found that compared the bitter-compound content of kawachi bankan with other cultivars by the same method.
• A way of removing only the bitter compounds without taking aroma and other compounds along with it has not yet been established.

Read in the encyclopedia

The encyclopedia entries below are in Japanese.

• limonin / naringin — the two compounds carrying delayed and immediate bitterness respectively
• orange juice / grapefruit juice — the juices studied for bitterness interplay and debittering
• grapefruit pith / pomelo pith — the part of the fruit high in naringin (limonoids are higher in the outer peel, membranes and seeds)
• Hassaku / Kawachi bankan / Citrus natsudaidai / amanatsu (Citrus natsudaidai) — Japanese mid-to-late citrus varieties that carry their own origin stories
• bitter orange marmalade / yuzu peel marmalade — how processing (salt water, high pressure) moves bitterness
• Japanese citrus beverages — the category where how bitterness is described splits from brand to brand

Go deeper in the research notes

• Citrus Bitterness Arrives Late — follows in more depth the immediate versus delayed bitterness, the differences in content between cultivars, and the gap in how the market talks about them
• Citrus Peel — Bitterness Left in Chenpi, Yuzu, and Marmalade — follows the bitterness left in the peel itself, through records from herbal medicine, processing, and standards
• Tonic Water and Quinine — A Drink That Puts a Ceiling on Bitterness — a comparison line for how a different lineage of bitterness, outside citrus, gets talked about in a beverage

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.