If you make coated candy in Europe, you have spent the last few years replacing titanium dioxide. Most of the replacements work. Very few of them look the same.
Phycocyanin does not replace titanium dioxide, and we sell phycocyanin. It cannot make anything white. What it might do is fix the specific thing that went wrong with the white you replaced it with, and that is worth a proper test.
Why the white changed
The European Commission withdrew titanium dioxide (E171) from the list of permitted food additives in Regulation (EU) 2022/63, after EFSA concluded in its 2021 safety assessment that it could no longer be considered safe as a food additive. Foods made under the old rules could be placed on the market until 7 August 2022. After that, the white in a sugar shell had to come from something else.
Titanium dioxide was hard to replace because of how it makes white. It bends light more strongly than almost anything you can eat, so a small amount scatters nearly all visible light straight back and hides whatever is under it. The usual stand ins scatter less. A patent filing on starch based replacements notes that rice starch or calcium carbonate need up to ten times the dose of titanium dioxide, which changes processing, appearance and crunch. Researchers working on titanium dioxide free tablet coatings note that calcium carbonate is commonly used in its place and that it gives a less white film.
Less scattering means the shell shows more of what is under and inside it: the slight yellow of the sugar or polyol, the brown of a chocolate centre, a xylitol coating that one emulsifier supplier describes as yellowish without titanium dioxide. The candy is still white. It just reads duller, and next to a pastel blue or pink in the same jar, the eye catches it.
What a blue can do to a white
A white that looks yellow is reflecting a little less blue light than the rest. Laundries solved this a century ago with bluing: a tiny amount of blue dye in the rinse. The blue absorbs some of the yellow side of the spectrum, the reflected light comes back closer to neutral, and the eye reads the result as cleaner and whiter. Plastics makers still do it: polyester gets a trace of blue toner to neutralise its yellow.
Phycocyanin absorbs strongly in the orange to red part of the spectrum, which is why it looks blue. At a trace, below the point where anyone would call the candy blue, it should pull a yellowish white toward neutral in the same way.
Anyone planning a reformulation should know its two limits.
It adds no opacity. Phycocyanin is a soluble protein, not a particle. It does nothing for hiding power. The opacifier still has to come from calcium carbonate, rice starch or whatever system you use today.
It makes the white slightly darker. Bluing works by taking light away, not adding it. An optical brightener in a detergent avoids that by glowing blue under UV, and phycocyanin cannot do that job: its own fluorescence is red. The colorimeter will show less yellow. What nobody has shown for candy, as far as we can find, is whether people prefer the neutral, slightly darker white to the brighter, yellower one. With laundry they do. That is what our trial is built to answer.
Where it can work and where it cannot
The shell has to be built cold. Phycocyanin starts to lose colour at high temperature, so it belongs in the panning syrup added at 50 °C or below, in compressed mints and in late dosed coatings. A boiled hard candy is cooked far too hot for it.
Below about pH 4.5 the colour fades and the protein can precipitate, which we cover in the low pH article. Calcium carbonate slurries sit on the other side, often around pH 8 to 9, and the trial below checks that end too.
Light is the bigger risk. A trace dose has the least colour to lose, so a pack that lets light in will show fading first. The light and packaging article has what we know about that.
On labelling: spirulina extract is sold in the EU as a colouring food, which we explain in the EU regulatory article, and the US lists it for candy and chewing gum, as covered in the US article.
The bench protocol
This is the protocol we are about to run. Use it, change it, or send us your white base and we will run it for you.
Matrices. Stage 1 is a quick screen: an opaque sugar syrup (60% sucrose, 30% water, 10% calcium carbonate or rice starch) cast as films in petri dishes and as drawdowns on black and white opacity cards. Stage 2 is a real hard panned shell on a centre, with the two or three best doses. Stage 3 stores the stage 2 pieces for 28 days in clear and opaque packs, in light and dark.
Doses. TintBlue E40, dosed from a fresh 1% cold solution into the white syrup at 50 °C or below, at 0, 5, 10, 20, 40, 80 and 160 ppm of product in the syrup. In the panned shell, that syrup goes on in the last white charges, the ones you see. For scale, our starting formula for a full blue panned shell puts 0.15% E40 (1,500 ppm) in the colour syrup, so the top of this ladder is about a tenth of a full blue and the bottom a three hundredth. Each step doubles the one before, which gives roughly even visual steps. If 5 ppm already reads blue, the ladder goes down; if 160 ppm still reads plain white, it goes up. TintBlue E18 does the same job at three times each dose.
pH. Measured on every syrup. If a calcium carbonate syrup reads above 8, a second arm is buffered down to about 7 to separate a pH effect from a dose effect.
Instrument readings. CIELAB under D65 with a 10° observer: L*, a*, b* and the CIE whiteness index, nine readings per dose in stage 1 and twenty in stage 2. Contrast ratio from the drawdowns, to confirm the blue does not change opacity. A titanium dioxide coated piece is read as the reference.
Panel. Eight or more people, coded samples on a grey card in daylight or a D65 booth. Two questions: rank the samples from whitest to least white, and, for each one, does it look blue?
Decision. A dose counts only if fewer than a quarter of the panel call it blue, the panel ranks it whiter than the undyed control, and L* drops by no more than 2. The idea passes if such a dose exists in a real panned shell and still holds after 28 days in an opaque pack. A better b* on the colorimeter, on its own, is not a result.
What we will publish
Whatever comes out, including a no. If it works, this page gets the dose, the matrix, the pack and the numbers. If it does not, it gets the reason. Results from anyone who runs it with us go up only with their yes.
What we will never write, whatever the result, is that phycocyanin replaces titanium dioxide, or that it is a natural whitener or an optical brightener. None of those is physically true.
If you are fighting a yellowish white in a coated sweet or a gum, write to benjamin@tint.bio, or take a free TintBlue sample (you pay the freight) and run the screen on your own base.
