Mineral sunscreen used to leave you grey. The chemistry did not change — the particle size did

There is a particular shade of chalky lavender-grey that anyone who has used an old mineral sunscreen will recognise instantly. It sat on the skin, refused to rub in, and announced itself from across a room. Cricket umpires and lifeguards wore it because it worked, and everybody else quietly bought something else.

The sunscreens sold today with the same two active ingredients — zinc oxide and titanium dioxide — can be almost invisible. The compounds are identical. The chemistry is identical. The only thing that changed is how finely the powder is divided, and following that single variable through explains the colour, the protection, the safety debate and most of what is on the label.

They do not reflect ultraviolet. They absorb it

The first thing to correct is the description almost every product page uses. Mineral filters are routinely called "physical blockers" that "sit on top of the skin and reflect UV like a mirror", in contrast to "chemical filters" that absorb it. That is a nice story and it is mostly wrong.

Zinc oxide and titanium dioxide are semiconductors, and like any semiconductor they have a band gap — an energy threshold below which a photon passes through and above which it is absorbed, promoting an electron across the gap. Zinc oxide's band gap is about 3.3 eV, which corresponds to light of roughly 375 nm. Titanium dioxide in its rutile form sits near 3.0 eV, anatase near 3.2 eV, in the region of 390 to 410 nm.

Ultraviolet light carries more energy than that threshold, so it is absorbed. Visible light carries less, so it is not. That single fact is why these two compounds are used at all: the cut-off falls almost exactly at the boundary between the ultraviolet you want stopped and the visible light you want to pass through so the skin still looks like skin.

The energy absorbed is dissipated mostly as heat, which is the same fate as the energy absorbed by an organic filter. For nanoscale particles, the great majority of ultraviolet attenuation is absorption; reflection and scattering contribute a minority share. The mirror image is a marketing metaphor that survived because it sounds reassuringly inert.

Their band-gap positions also differ usefully. Zinc oxide's cut-off extends further into the long-wavelength UVA1 region, near 370–380 nm, which is why it is prized for broad-spectrum protection. Titanium dioxide is stronger in the UVB region and short UVA. Many formulations use both.

Why the old ones were white, and the new ones are not

If ultraviolet absorption is set by the band gap, and the band gap does not care about particle size, why does size change anything at all?

Because absorption is not the only thing happening. Particles also scatter light, and scattering depends very steeply on the relationship between particle size and wavelength.

When a particle is much smaller than the wavelength of the light hitting it, scattering is weak and falls away extremely fast as the particle shrinks — in the Rayleigh regime it scales with the sixth power of diameter. When the particle is comparable to the wavelength, scattering is strong and broadly white. Visible light runs from about 400 to 700 nm, so a pigment-grade oxide particle of 200 to 300 nm sits right in the range that scatters visible light powerfully. That is not a coincidence: the same materials, at that same size, are the world's dominant white pigments, in paint, paper, toothpaste and the icing on a cake.

Now shrink the particle to 30 to 50 nm. It is now far smaller than any visible wavelength, so visible scattering collapses by orders of magnitude and the film turns transparent. But ultraviolet at 300–380 nm is still absorbed, because absorption is a band-gap property and the band gap is unchanged. The white cast disappears and the protection stays.

That is the whole trick, and it is one of the cleanest illustrations available of why size behaves like a material property at the nanoscale.

It is not free, though. Two things push back as particles get smaller. Very small zinc oxide particles lose some effectiveness at the long-UVA end, so a formulator chasing perfect transparency can quietly give up the protection that made zinc oxide worth using. And small particles are strongly inclined to agglomerate — clumps behave optically like the big particles you were trying to avoid, so dispersion stability in the finished emulsion decides whether the transparency survives to the end of the tube. What looks like a formulation aesthetics problem is really a colloid-stability problem.

The property that had to be switched off

Here is the part that makes these two oxides genuinely interesting rather than merely useful.

Titanium dioxide under ultraviolet light is a photocatalyst. The electron promoted across the band gap, and the positively charged hole it leaves behind, can migrate to the particle surface and drive reactions there — generating reactive oxygen species from water and oxygen. Zinc oxide does the same. This is not a defect. It is a celebrated and commercially valuable property: it is precisely the mechanism behind self-cleaning glass and photocatalytic coatings, where the whole point is to generate radicals that break down grime.

On a building, that is the feature. On skin, it is the last thing you want, and it is worse than merely unhelpful — free radicals generated at the surface of the particle sit directly against the substrate you applied the sunscreen to protect.

So sunscreen-grade oxides are not raw oxides. They are surface-treated: coated with a thin shell of silica, alumina, or a hydrophobic layer such as dimethicone or a stearate, sometimes several of these in combination. The coating does two jobs. It puts a barrier between the photogenerated radicals and everything outside the particle, suppressing surface photoactivity by a large margin. And it controls how the particle disperses in the oil or water phase of the emulsion, which is the agglomeration problem from the previous section.

Crystal phase matters here too. Rutile titanium dioxide is markedly less photoactive than anatase, which is why sunscreen grades are rutile and photocatalytic self-cleaning grades are anatase, or an anatase-rich mixture. Two products, the same compound, deliberately opposite specifications.

The same property is a selling point in one product and a defect in another. This is the ordinary condition of materials science, and it is why a compound name is never a specification — the phase, the particle size and the surface treatment carry the actual engineering.

What SPF actually promises, and what it does not

Given a well-made filter, the number on the front of the bottle is the next thing worth reading properly.

SPF measures UVB protection, because it is defined by sunburn — the ratio of the ultraviolet dose needed to redden protected skin to the dose needed to redden unprotected skin. It is measured on human volunteers, and it is measured at an application density of 2 mg/cm², which is roughly a heaped teaspoon for a face and neck and about 30 ml for a whole body.

Almost nobody applies that much. Real-world use is commonly a quarter to a half of it, and the protection does not fall proportionately — it falls faster, because attenuation is exponential in film thickness. The practical consequence is that a high number applied thinly can deliver less than a moderate number applied properly, and that reapplication matters more than the printed figure.

The numbers themselves also compress badly at the top, in the same way that antimicrobial log-reduction figures do. SPF 15 attenuates roughly 93% of erythemal UVB, SPF 30 about 96.7%, SPF 50 about 98%. Going from 30 to 50 is a real improvement, but it is a much smaller step than the arithmetic suggests, and the marketing gap between 50 and "50+" is smaller still.

UVA is a separate measurement, and this is where labels differ by region. UVA does not burn efficiently but penetrates deeper and drives photoageing and DNA damage. The European approach requires a critical wavelength of at least 370 nm and a UVA protection factor of at least one third of the labelled SPF, shown as the circled "UVA" mark. The Japanese-origin PA+ to PA++++ system, widely used in India and across Asia, grades persistent pigment darkening instead. A high SPF with no UVA marking is an incomplete product, whatever the number says.

What the safety evidence actually shows

The nano-in-sunscreen argument has been running for two decades, and the evidence is now reasonably settled on the main question, while remaining genuinely open on a smaller one.

Skin penetration has been studied repeatedly, including on volunteers, on damaged and sunburned skin, and with isotopically labelled zinc. The consistent finding is that these particles stay in the stratum corneum — the dead outer layer — and in follicular openings, and do not reach living epidermis in meaningful quantity. The isotope work did detect a very small systemic uptake of zinc, but as dissolved ions rather than as particles, at levels small against normal dietary zinc. The European Commission's scientific committee has repeatedly reviewed both oxides for use in cosmetics on this basis.

Inhalation is the exposure route that genuinely warrants care, and it is a different question from skin contact entirely. Inhaled titanium dioxide is classified by IARC as possibly carcinogenic to humans (Group 2B) on inhalation evidence in animals — a classification about breathing the powder, not about wearing it. The practical translation is specific and worth acting on: spray and loose-powder mineral sunscreens create an inhalable aerosol, and creams and lotions do not. Avoid spraying near the face; there is no reason to accept the one exposure route the evidence actually flags.

Environmental questions are the ones still open. The reef-protection laws passed in Hawaii, Palau and elsewhere target the organic filters oxybenzone and octinoxate, not mineral filters — mineral sunscreens are usually the substitute rather than the target. But nano zinc oxide and titanium dioxide entering coastal water in large volumes have their own ecotoxicology, including dissolved zinc toxicity to marine organisms, and calling a mineral sunscreen "reef safe" is a marketing claim rather than a regulated one.

Reading a metal-oxide specification

For anyone formulating with these materials or working on them in a laboratory, "zinc oxide" and "titanium dioxide" describe a compound, not a grade. Five entries on the datasheet decide behaviour.

1. Crystal phase. Rutile or anatase for titanium dioxide, and it changes photoactivity by a large factor. Specify it; do not infer it.
2. Primary particle size and the distribution around it. A mean of 40 nm with a long tail into the hundreds behaves differently from a tight distribution, and the tail is what you see as haze.
3. Surface treatment. Which coating, and how much of it. Silica, alumina, stearate, silicone or a stack; this governs both photoactivity suppression and dispersibility, and it also changes what fraction of the powder is actually active oxide.
4. Specific surface area, by BET. For particles you cannot easily image, this is the practical size proxy, and it is directly comparable between suppliers.
5. Purity and residual metals. Heavy metal limits are a regulatory requirement for cosmetic-grade material and a confounder in toxicology work, and this is the line that separates a cosmetic grade from an industrial pigment grade at a similar nominal size.

Dispersion protocol belongs alongside these rather than after them. A well-specified powder that agglomerates in your emulsion will underperform a modest grade that stays dispersed, and pre-dispersed grades exist for exactly that reason.

Why it matters for students and researchers

This is an unusually good teaching example because a single variable — particle size — visibly controls an optical property, a functional property and a regulatory argument at the same time, and the student can see each one without any apparatus more exotic than their own eyes.

It also demonstrates a habit worth acquiring early: reading past the compound name. Rutile and anatase titanium dioxide are chemically identical and functionally opposite for this purpose. A coated and an uncoated particle of the same size have very different surface chemistry. Anyone who learns to ask "which phase, what size, what coating" before accepting a claim about a material has learned most of what characterisation is for.

The open problems are real ones. Long-term environmental fate of coated metal-oxide nanoparticles, including whether coatings persist through weathering and wastewater treatment, is not settled. Coating durability under prolonged solar exposure — whether the photoactivity suppression measured on fresh material still holds after hours on skin in the sun — deserves more published work than it has. And the gap between laboratory SPF measured at 2 mg/cm² and what people actually achieve is a behavioural and formulation question with a larger public-health effect than another increment of SPF.

Frequently asked questions

Is mineral sunscreen better than chemical sunscreen?

They work by more similar mechanisms than the labels suggest — both absorb ultraviolet, and mineral filters simply do it via a semiconductor band gap rather than an organic chromophore. Mineral filters are photostable and less likely to cause contact sensitivity, which makes them the usual recommendation for sensitive and infant skin; organic filters are easier to formulate into an elegant, high-SPF, genuinely invisible product. The best sunscreen remains the one that is applied thickly and reapplied, and the choice between the two families matters much less than that.

Are nanoparticles in sunscreen absorbed into the body?

The evidence says essentially no, for intact and for sunburned skin. Studies consistently find the particles confined to the stratum corneum and hair follicle openings. Isotope-labelled zinc studies found trace systemic zinc, absorbed as dissolved ions rather than as particles, at levels small compared with dietary intake. The exposure route that does warrant caution is inhalation, not skin, which is an argument against spray and powder formats specifically.

Why do some mineral sunscreens still leave a white cast?

Either the particles are deliberately not nanoscale, or they have agglomerated. Some brands use non-nano grades in response to consumer preference and accept the cast; some tinted formulations add iron oxides to mask it. And in any formulation, poor dispersion lets small particles clump into optically large ones, which scatter visible light and bring back the very whiteness the small particle size was chosen to avoid.

If titanium dioxide breaks down dirt on self-cleaning glass, what is it doing on my face?

On glass it is anatase and uncoated, because photocatalysis is the product. In sunscreen it is rutile and surface-coated with silica, alumina or a silicone layer, precisely so that photocatalysis does not happen. Same compound, opposite specification, chosen deliberately in each case.

Does a higher SPF mean proportionally more protection?

No — the scale flattens quickly. SPF 30 attenuates about 96.7% of erythemal UVB and SPF 50 about 98%, so the second half of that jump buys far less than the number implies. How much you apply matters more: laboratory SPF is measured at 2 mg/cm², most people apply a fraction of that, and protection falls faster than the shortfall. Check for a UVA mark too, since SPF alone says nothing about it.