Butterfly Pea: Anthocyanins and Why It Changes Colour

Pour a cup of butterfly-pea tea and you get an ink blue that looks like it must be artificial. Squeeze in lemon and it slides through violet to pink in a couple of seconds. Nobody who watches this happen for the first time quite believes it came out of a flower.

It did, and the explanation is one of the more satisfying pieces of everyday chemistry available to a home kitchen. It is not a dye reaction, not oxidation, and not the pigment "breaking down". It is the same molecule sitting in four different structural forms, whose relative abundance is set by how many free protons are floating around it. Add acid, shift the balance, change the colour. Take the acid away and it comes back. This is the definition of an acid–base indicator, and butterfly pea is a good one.

This page explains that properly, because it is the most honest interesting thing about the plant. It also draws a line that the marketing works hard to smudge: a colour change tells you the pH of the liquid, and absolutely nothing about what the liquid does to your body. Litmus paper turns red in acid too, and nobody sells it as a supplement.

Table of Contents

1. Why This Page Starts With Chemistry
2. What an Anthocyanin Actually Is
3. Delphinidin: The Blue End of the Family
4. Ternatins and the Armour of Polyacylation
5. The Four Forms and the pH Ladder
6. Running the Demonstration at Home
7. Why a Colour Change Is Not a Health Claim
8. Heat, Light, Oxygen and Metals
9. Why Industry Wants a Natural Blue
10. Compared With Other Anthocyanin Sources
11. Brewing for Colour
12. Key Research Papers
13. Connections

Why This Page Starts With Chemistry

Most herb pages open with what the plant is supposed to do for you. This one opens with molecular structure, deliberately, because with butterfly pea the structure is the story and everything else is downstream of it.

Consider what actually happened commercially. Clitoria ternatea had been growing on fences across tropical Asia for centuries and colouring rice in Kelantan and Bangkok the whole time. It became a global ingredient in about a decade, and it did so because of a single visual property that photographs and films beautifully. No health claim drove that. The colour did. Then the health claims arrived to monetise the attention the colour had already earned.

So if you understand the colour, you understand where the hype came from and why it points in the direction it does. You also acquire the tool you need to evaluate the rest: the difference between a property of a molecule in a glass and an effect on a person.

What an Anthocyanin Actually Is

Anthocyanins are the pigments responsible for most red, purple and blue colour in the plant kingdom — in blackcurrants, red cabbage, aubergine skin, autumn leaves, red wine and blue flowers. They belong to the flavonoids, the same broad family as the tea catechins and the citrus flavanones.

Every anthocyanin has three parts, and each part matters for the colour:

1. The aglycone core, called an anthocyanidin. This is a flat, three-ringed structure — two aromatic rings joined by an oxygen-containing central ring. In its coloured state it carries a positive charge and is called a flavylium cation. Only six anthocyanidins matter in food plants: pelargonidin, cyanidin, delphinidin, peonidin, petunidin and malvidin. They differ only in how many hydroxyl and methoxyl groups decorate one ring, and that small difference sets the base hue: pelargonidin runs orange-red, cyanidin red-magenta, delphinidin blue-purple.
2. Sugars, attached by glycosylation. Anthocyanidins are unstable naked; in living tissue they are almost always glycosylated, most often with glucose, at the 3-position and frequently elsewhere too. The sugars improve water solubility and stability. An anthocyanidin plus its sugars is an anthocyanin.
3. Acyl groups, attached by acylation. Some anthocyanins carry organic acids — p-coumaric, caffeic, ferulic, malonic — esterified onto the sugars. This is the optional extra that transforms everything, and butterfly pea has more of it than almost anything else you will eat.

The reason this matters for a drink is that an unacylated anthocyanin is a fragile pigment. Cyanidin-3-glucoside, the workhorse anthocyanin of blackberries and blackcurrants, is vividly coloured in acid and fades to a washed-out grey-lilac within minutes as pH approaches neutral. That is why almost every anthocyanin-coloured food product is acidic, and why blue is so rare on the natural-colour shelf.

Delphinidin: The Blue End of the Family

Butterfly pea's pigments are built on delphinidin, the most hydroxylated of the six common anthocyanidins, carrying three hydroxyl groups on its B-ring. Extra hydroxylation pushes the absorbance maximum to longer wavelengths, which shifts the perceived colour from red toward blue. Delphinidin is why delphiniums, violas, and the blue morning glory are blue, and it is the necessary starting point for any plant attempting a true blue.

Necessary but not sufficient. Delphinidin on its own is also the least stable of the anthocyanidins — the same electron-rich B-ring that gives the blue shift makes it easier to oxidise. Plants that succeed at blue have to solve stability separately, and evolution has found several answers: raising the vacuolar pH inside the petal cell, chelating the pigment to aluminium or iron ions, stacking it with colourless copigment flavonoids, or building protective acyl groups into the molecule itself.

Hydrangeas use the metal route, which is why soil aluminium availability flips them between pink and blue. Cornflowers use a large pigment–metal–copigment complex. Butterfly pea uses the molecular-armour route, and does it more thoroughly than nearly any other edible plant.

Ternatins and the Armour of Polyacylation

The butterfly-pea pigments are called ternatins, after the species epithet. Norio Terahara's group in Japan isolated and characterised them across a series of papers in the 1990s, naming the ternatin A, B, C and D series — plus "preternatins" found in young, not-yet-open flowers.

Structurally, each ternatin is delphinidin glycosylated at the 3, 3′ and 5′ positions, with two long side chains extending from the 3′ and 5′ glucoses. Each side chain is built from alternating glucose and p-coumaric acid units, like beads on a string. The largest members of the series carry four coumaroyl groups on each of the two chains. That is an extraordinary amount of decoration for a food pigment; most acylated anthocyanins in the diet carry one or two acyl groups in total.

What those chains do is physical, not chemical. They are flexible, and they are flat aromatic rings on flexible tethers. Left to themselves in water they fold back over the flavylium core and stack face-to-face with it, held there by the same weak π-stacking forces that hold the rungs of DNA together. The effect is a molecular sandwich: the reactive flat face of the pigment is physically covered by its own side chains.

This is called intramolecular copigmentation, and its consequence is precisely what butterfly pea is famous for. The main route by which an anthocyanin loses its colour at neutral pH is nucleophilic attack by a water molecule at the pigment's 2-position, converting the coloured flavylium into a colourless hemiketal. If the 2-position is buried under a stack of coumaroyl rings, water cannot easily get to it. The pigment therefore stays coloured in conditions where an ordinary anthocyanin would have gone grey.

Two useful corollaries follow, and both are testable in a kitchen:

• Butterfly pea holds blue in near-neutral water for hours, where a red-cabbage extract at the same pH turns a murky blue-grey much faster. The armour is doing real work.
• The armour is not invincible. Heat gives the side chains enough energy to unstack, and prolonged light exposure degrades the pigment outright. Butterfly-pea colour is stable relative to other anthocyanins, not stable in absolute terms.

Alongside the ternatins the petal contains colourless flavonol glycosides — kaempferol, quercetin and myricetin derivatives, including malonylated forms characterised by Kazuma and colleagues. These can act as intermolecular copigments, stacking with the pigment from outside and adding a further increment of blue stabilisation. The petal is, in effect, running two copigmentation strategies at once.

The Four Forms and the pH Ladder

Here is the part that most explanations skip, and it is the part that makes the colour change comprehensible rather than magical.

An anthocyanin in water is not one species. It is an equilibrium mixture of four interconverting forms, and pH determines which dominates:

1. The flavylium cation — the protonated, positively charged form. Red to magenta. Dominant in strong acid.
2. The quinoidal base — formed by losing a proton from a hydroxyl group, giving a neutral molecule with an extended conjugated system. Blue to violet. Dominant from mildly acidic to neutral. Butterfly pea's whole reputation rests on this form surviving there.
3. The carbinol pseudobase, or hemiketal — formed when water adds across the 2-position. Colourless. This is the form that ruins ordinary anthocyanins at neutral pH, and the one the ternatin side chains are shielding against.
4. The chalcone — formed when the hemiketal's central ring opens up. Pale yellow, and the slowest form to appear and disappear. In alkali, and after prolonged heating, this is where the pigment ends up.

Now read the colour sequence as a ladder up the pH scale. Approximate values, because the exact transition points depend on the specific ternatin, the temperature and what else is dissolved in the cup:

• Around pH 1–2 — strong acid. Flavylium cation dominates. Bright pink-magenta.
• Around pH 3–4 — lemon or lime juice territory. A mixture of flavylium and quinoidal base. Purple, and the exact shade depends on how much citrus you added, which is why the transition looks gradual and controllable.
• Around pH 5–7 — plain drinking water, roughly neutral. Quinoidal base dominates. The classic butterfly-pea blue.
• Around pH 8–9 — a pinch of baking soda. Deprotonation continues and the balance shifts toward greener hues; a blue-green or teal.
• Above about pH 10 — strong alkali. The ring opens toward the chalcone. Yellow-green to yellow-brown.

Two features of this deserve emphasis, because they are the details that make the demonstration convincing:

It is reversible, up to a point. The flavylium↔quinoidal transition is a proton transfer and is essentially instantaneous and fully reversible — add acid, get pink; neutralise it, get blue back. That reversibility is what proves it is an equilibrium shift rather than destruction. The hemiketal and chalcone steps involve making and breaking covalent bonds, are much slower, and are only partly reversible; once you have cooked a batch to yellow-brown in alkali, it is not coming all the way back.

The colour is a genuine measurement. Because the transition tracks proton concentration, butterfly-pea extract functions as a real, if coarse, pH indicator, and has been used as one in teaching laboratories and in food-freshness indicator films — edible packaging that changes colour as a product spoils and its pH drifts. That is a legitimate technical application of the pigment, and worth knowing about because it is the clearest possible illustration of what the colour change actually reports on: the liquid, not the drinker.

Running the Demonstration at Home

This is one of the best kitchen-science demonstrations available, it costs almost nothing, and everything in it is food. Tier: this is a physical demonstration, not a health intervention.

1. Steep eight to ten dried butterfly-pea flowers in about 250 mL of just-boiled water for five minutes, then strain. You want a deep blue you cannot see through at the bottom of the cup.
2. Divide it between several clear glasses — clear glass matters, since the whole point is watching.
3. Leave the first glass alone. That is your reference blue at roughly neutral pH.
4. Add a few drops of lemon or lime juice to the second and swirl. Violet, within a second or two.
5. Add a full teaspoon or more of citrus juice to the third. Pink to magenta.
6. Add a small pinch of baking soda to the fourth. It fizzes slightly and turns green-teal, then muddies toward yellow-brown if you add more.
7. Now go back to the pink glass and stir in a very small pinch of baking soda. It travels back through violet toward blue. That reversal is the proof that you shifted an equilibrium rather than destroyed a pigment.

Practical notes from getting this wrong:

• Add citrus at the table, not during the steep. Acid in the brewing water means you never see blue at all — you get purple tea from the start and no transition to show anyone.
• Hard or alkaline tap water shifts your starting point. If your reference glass looks green-blue rather than blue, your water is more alkaline than you think. Use filtered or bottled water for the demonstration.
• Sparkling water pulls it purple — dissolved carbon dioxide forms carbonic acid. This is also why butterfly pea in a carbonated drink never looks as blue as it did in the jug.
• Do not use it to make a claim. If a child asks what the colour change means, the true and much more interesting answer is "it means lemon juice is acidic and water is not", not "it means the tea is good for your brain".

Why a Colour Change Is Not a Health Claim

This is the section the rest of the page exists to support, so it is worth being direct.

The colour change is a pH indicator response. It is not evidence of any biological effect, and it cannot be. The reaction you are watching happens in a glass, at a proton concentration you controlled, in a molecule that has not been anywhere near a cell. Litmus, phenolphthalein, red cabbage and turmeric all do versions of the same thing. Their colour changes are equally real and equally silent on the subject of human health.

Several specific leaps show up in butterfly-pea marketing, and each one is worth naming:

• "It changes colour, so it is bioactive." Changing colour with pH is a property of a class of pigments so common that it is a school experiment. It has no relationship to whether a compound reaches your bloodstream or does anything there.
• "It changes colour in your stomach, so it is working." Stomach acid is around pH 1–2, so yes, ternatins turn pink there. So does the anthocyanin in every blackberry you have ever eaten. This tells you about the stomach, not about the pigment's fate — and the actual fate, covered on the antioxidant and metabolic claims page, is that most of it is never absorbed intact at all.
• "The blue means it is high in antioxidants." Deep colour does correlate loosely with anthocyanin content, and anthocyanins do register on test-tube antioxidant assays. But the amount of pigment in a cup of flower infusion is small, the assay is a chemistry measurement rather than a health outcome, and the withdrawn USDA ORAC database is the cautionary tale for exactly this reasoning.
• "It detoxifies because it neutralises acid." The pigment does not neutralise anything — it responds to acid. It is the thermometer, not the heater. And blood pH is held within a range of a few hundredths of a unit by respiratory and renal control; no drink shifts it, and you would not want one that did.

None of this is an argument against drinking butterfly-pea tea. It is an argument for enjoying it as what it is: a striking, caffeine-free, almost flavour-neutral infusion with a genuinely excellent piece of chemistry in it. The chemistry is the benefit on this particular page. It is a real one, and it does not need to be inflated.

Heat, Light, Oxygen and Metals

If you have ever wondered why your dried flowers went from indigo to a sad grey-blue in the cupboard, or why a homemade blue syrup turned brown, this is why. Four things degrade ternatins, and understanding them is the practical payoff of the structural story above.

• Heat. Prolonged heating drives the equilibrium toward the colourless hemiketal and the yellow chalcone, and gives the protective side chains enough thermal energy to unstack from the core. Brief exposure to just-boiled water is fine — that is how you brew it. Simmering a syrup for twenty minutes is not.
• Light. Anthocyanins are photodegraded, and the flat aromatic pigment absorbs visible light by definition. Dried flowers stored in a clear jar on a sunny shelf fade measurably over months. Store them in an opaque or dark container.
• Oxygen, and oxidative enzymes. Air exposure combined with residual plant enzymes and ascorbic acid can accelerate degradation. This is one reason a fresh infusion looks better than one left open overnight, and why vitamin C added to an anthocyanin drink can, counter-intuitively, hasten pigment loss.
• Metal ions. Iron and aluminium chelate delphinidin-type pigments and shift the hue — sometimes toward a stable blue, which some plants exploit deliberately, and sometimes toward an unappetising grey. Brewing in an iron or unlined aluminium vessel can change your colour unpredictably. Glass, ceramic or stainless steel are safe.

The storage rules follow directly: dried flowers in an airtight, opaque container, in a cool dark cupboard, and judge quality by colour — good stock is a deep indigo-blue, and a faded grey-blue means old or light-damaged flowers that will brew weakly.

Why Industry Wants a Natural Blue

There is a commercial context to all of this that explains why so much research money has gone into butterfly-pea chemistry specifically, and it makes the plant's story much clearer.

Natural blue is the hardest colour in the food industry. Manufacturers replacing synthetic dyes can find natural reds, oranges, yellows, greens and browns without much difficulty — beet, paprika, annatto, turmeric, carotenoids, chlorophyll, caramel. Blue is a near-vacuum. The candidates are few and each has a serious drawback: spirulina-derived phycocyanin is a protein and is heat- and acid-sensitive; the genistein-based blues are novel and expensive; and ordinary anthocyanins simply will not hold blue outside a narrow pH window.

Butterfly pea walked into that gap with three advantages at once. Its pigment holds a true blue in near-neutral products because of the polyacylation described above. Its flavour is close to neutral — faintly grassy, faintly like a mild green tea — so it colours without tasting, which is a strict requirement for a colourant. And it has a long culinary history in Southeast Asia, which matters enormously to a regulatory dossier.

That last point is why the plant's regulatory status is worth checking rather than assuming. Butterfly-pea flower extract has been evaluated as a colour additive for use in certain foods and beverages in the United States, and permitted categories have been expanded over time; the position differs by jurisdiction and by food category. If this matters to you commercially, read the current regulator position rather than a supplier's summary.

The same stability that makes it a colourant also makes it useful in intelligent packaging: films loaded with butterfly-pea extract have been studied as freshness indicators, changing colour as a packaged food's pH drifts during spoilage. It is a nice closing of the loop — the property that made it a novelty drink is being put to work as a sensor.

Compared With Other Anthocyanin Sources

Placing butterfly pea beside its neighbours makes both its genuine distinction and its genuine limitation obvious.

• Red cabbage is the other classic pH-indicator vegetable, and its anthocyanins are also acylated, which is why it manages a blue-purple rather than collapsing entirely at neutral pH. It runs the same colour ladder. Its drawbacks as a colourant are flavour and aroma — it tastes like cabbage, which butterfly pea does not.
• Hibiscus is the great red flower tea, and its anthocyanins are largely unacylated delphinidin and cyanidin sambubiosides. It is vividly red because the infusion is genuinely acidic — the tartness and the colour come from the same place. Hibiscus cannot be blue, and butterfly pea cannot be tart. They are frequently blended, and the blend always leans purple, because the hibiscus acid pulls the ternatins down the pH ladder.
• Bilberry and blueberry are the anthocyanin-density champions of this comparison. Per gram of edible material they deliver far more anthocyanin than a flower infusion does, which is the central reason a cup of blue tea is a modest polyphenol source in absolute terms — a point developed on the antioxidant claims page.
• Grapes and red wine supply malvidin-based anthocyanins, some acylated, and are the most studied dietary anthocyanin source of all. They also demonstrate the copigmentation principle in a way wine drinkers already know: a young wine's colour depends on pigment–tannin copigment complexes, and it shifts as those complexes evolve.

The honest placement, then: butterfly pea is genuinely unusual for the stability and hue of its pigment, and genuinely unremarkable for the quantity delivered in a normal serving. Those two facts are not in tension; they are simply about different things. A pigment can be the best in its class at being blue and still be present in a cup at a few milligrams.

Brewing for Colour

Everything above collapses into a short set of practical rules.

1. Water just off the boil, five minutes, eight to ten dried flowers per 250 mL. Longer steeping extracts more but also more of the faint grassy note; it does not deepen the blue much beyond about five minutes.
2. Use neutral water. Filtered or bottled if your tap water is hard. Your starting colour is your reference point, and alkaline water starts you green.
3. Acid last, at the table. The show is the transition, and you only get it once per glass.
4. Cold-brew works and preserves colour better. Flowers in cold water in the refrigerator for several hours gives a clean blue with less thermal degradation. Good for large batches and for iced service.
5. Sweeten with what you like, but note that honey is mildly acidic and will nudge the colour toward violet. Sugar is neutral and will not.
6. Serve in glass. This is the one drink where the vessel is part of the point, and metal can shift the hue.
7. It is caffeine-free, which is a real and often-overlooked practical benefit: it gives you a visually interesting drink for the evening without a stimulant. That is a legitimate reason to keep it in the cupboard.

Key Research Papers

Cited as PubMed topic searches rather than as direct numeric identifiers. Where a numeric detail or exact metadata is not certain, the finding is described and the search link provided so you can read the original yourself.

1. Terahara and colleagues, Journal of Natural Products, 1996 — "Five new anthocyanins, ternatins A3, B4, B3, B2 and D2, from Clitoria ternatea flowers." The founding structural paper of the ternatin series. Search PubMed.
2. Terahara and colleagues, Journal of Natural Products, 1998 — "Eight new anthocyanins, ternatins C1–C5 and D3 and preternatins A3 and C4 from young Clitoria ternatea flowers." Establishes the immature-flower pigments. Search PubMed.
3. Kazuma, Noda and Suzuki, Phytochemistry, 2003 — malonylated flavonol glycosides from Clitoria ternatea petals: the colourless copigments that sit alongside the ternatins. Search PubMed.
4. Yoshida, Mori and Kondo, Natural Product Reports, 2009 — blue flower colour development by anthocyanins, from chemical structure to cell physiology. The clearest account of why blue is chemically hard and of the four strategies plants use. Search PubMed.
5. Vidana Gamage, Lim and Choo, Frontiers in Plant Science, 2021 — a review of butterfly-pea anthocyanin biosynthesis, extraction, stability and applications; the current synthesis of the stability literature. Search PubMed.
6. Oguis, Gilding, Jackson and Craik, Frontiers in Plant Science, 2019 — the comprehensive species review, covering pigment chemistry, cyclotides and agricultural use in one place. Search PubMed.
7. On the underlying equilibrium chemistry — the four-form flavylium / quinoidal base / hemiketal / chalcone system and its pH dependence is standard anthocyanin physical chemistry, developed largely in the 1970s and 1980s. Search PubMed.
8. On intramolecular copigmentation as a stabilisation mechanism in polyacylated anthocyanins — the mechanism that makes ternatins unusual. Search PubMed.
9. On butterfly-pea extract in colorimetric freshness-indicator films and intelligent food packaging — a growing applied literature that uses the pH response as a sensor. Search PubMed.
10. On the thermal and photostability of butterfly-pea colourant in real beverage systems — the practical literature behind the storage advice above. Search PubMed.
11. On copigmentation between anthocyanins and flavonols in flower petals — the intermolecular half of butterfly pea's two-strategy approach. Search PubMed.
12. On metal chelation and blue flower colour, including the hydrangea aluminium case — the alternative solution butterfly pea did not use. Search PubMed.

Connections

• All Herbs
• Butterfly Pea (Clitoria ternatea) — the main topic page, with botany, names, culinary use and dosage.
• Antioxidant and Metabolic Claims Examined — what happens to these pigments after you swallow them, and why that matters more than the colour.
• Anthocyanins — the full pigment class: sources, chemistry and the wider human evidence.
• Anthocyanins — Benefits — where the anthocyanin evidence base is genuinely strong, and where it is not.
• Hibiscus — the red flower tea whose colour and tartness come from the same acidity.
• Bilberry — a far denser anthocyanin source, and the classic case of anthocyanin overclaiming.
• Blueberries — more anthocyanin per gram than any flower infusion.
• Blackberries — cyanidin-3-glucoside, the unacylated anthocyanin that fades where ternatins hold.
• Cabbage — red cabbage is the other great kitchen pH indicator, and for the same structural reason.
• Quercetin — one of the copigment flavonols present in butterfly-pea petals.
• Lemongrass — the most common blending partner in Southeast Asian butterfly-pea drinks.
• Pandan — the other great Southeast Asian colouring and flavouring leaf, used alongside bunga telang in Nyonya sweets.

Safety and disclaimer. This page describes chemistry, not treatment, and is educational rather than medical advice. The colour demonstration uses only food ingredients and is safe, including for children, but nothing about a colour change indicates a health effect. Butterfly-pea flowers as a tea or colouring have a long culinary record with no established toxicity signal; concentrated extracts have not been studied for long-term human use. Avoid concentrated extracts in pregnancy and breastfeeding, and note that the seeds and root — traditionally purgative — are not the same material as the flower. Discuss any supplement with your clinician if you take prescribed medication, particularly for blood glucose.

Back to Table of Contents