Ingredients

Mineral or chemical sunscreen filters: how they work

Mineral filters do not bounce the sun away. Measured in an optical integrating sphere, zinc oxide and titanium dioxide reflect an average of 4 to 5 percent of the ultraviolet reaching them, less than an SPF of 2, and absorb almost all the rest. The differences that matter sit elsewhere: the European list permits about 30 filter substances at limits from 3 to 25 percent, the American monograph carries 16, and only the organic filters have been measured in the FDA's blood trials.

What a filter does to a photon

Sunscreen filters come in two families. The organic ones, usually called chemical, are carbon based molecules whose ring systems take in an ultraviolet photon and release the energy mostly as heat. The inorganic ones are two metal oxides, zinc oxide and titanium dioxide, sold as physical or mineral filters that supposedly sit on the surface and bounce the sun away.

That claim has been measured. In an integrating sphere, which collects light leaving a sample in every direction, the average reflection of zinc oxide and titanium dioxide across the ultraviolet range was 4 to 5 percent, less than an SPF of 2. The rest came from semiconductor band gap absorbance: the oxide takes the photon in, exactly as an organic filter does.

Reflection is real, but at the wrong end of the spectrum. Past the band gap wavelength, in the long UVA and visible range, the oxides stop absorbing and reflect up to 60 percent of the light, which is the physics of the white cast. It is also why the American definition of a sunscreen active ingredient covers absorbing, reflecting and scattering across 290 to 400 nm.

Two rule books and two lists

Europe and the United States treat filters as different kinds of product. Annex VI of the EU cosmetics regulation runs to entry 34 and permits about 30 substances, four of which appear twice because the nanomaterial form carries its own conditions. Limits run from 3 percent at the bottom of the list to 25 percent for both oxides, with avobenzone at 5 percent, ethylhexyl methoxycinnamate at 10 percent, octocrylene at 10 percent (9 percent in propellant sprays) and drometrizole trisiloxane at 15 percent.

Two amendments in 2022 tightened the older organic filters on the advice of the Commission's consumer safety committee. Benzophenone-3, known in America as oxybenzone, was cut to 6 percent in face, hand and lip products, 2.2 percent in body products and 0.5 percent elsewhere. Homosalate was cut to 7.34 percent and to face products other than propellant sprays, with non-compliant stock off the EU market from 1 January 2025 and out of shops from 1 July 2025.

The American list is shorter. 16 active ingredients sit in the monograph stayed since 1999, and eligibility for it was originally limited to ingredients already sold in the country when the review opened in 1972, which is why no new filter has joined the list since the late 1990s. In February 2019 the FDA proposed that only zinc oxide and titanium dioxide, at up to 25 percent, are generally recognised as safe and effective; that aminobenzoic acid and trolamine salicylate are not, for allergic reactions and for the clotting effects of salicylic acid; and that the other 12, avobenzone and oxybenzone among them, need more data. The same document proposed capping the label at SPF 60+ with formulation to SPF 80.

The UVA half of the label

UVB, 290 to 320 nm, does most of the burning and is what SPF measures, while UVA runs on to 400 nm, reaches deeper and drives pigment change and photoageing. An SPF of 50 says nothing about that longer half, so both systems add a second measure.

In the United States the words broad spectrum are earned by a critical wavelength of 370 nm or more, the point at which 90 percent of the area under the absorbance curve has been reached. The 2019 proposal would add a UVA I to total UV ratio of 0.7 or higher, because a product can pass the 370 nm test and still be weighted towards UVB. Europe's version is a recommendation, not a rule: the 2006 Commission recommendation asks for a UVA protection factor of at least a third of the labelled SPF and a critical wavelength of at least 370 nm, and the letters UVA in a circle are the industry mark for meeting it, measured in vitro by ISO 24443.

Avobenzone is the only strong UVA I absorber in the American monograph and it breaks down in sunlight, losing 50 to 90 percent of its effect after 60 minutes, so it has to be paired with a photostabiliser. The European list carries photostable alternatives the monograph does not: bemotrizinol and bisoctrizole at up to 10 percent, drometrizole trisiloxane at 15 percent. Zinc oxide covers UVA alone but returns little SPF per percent used, which is why mineral formulas sit near the 25 percent ceiling.

What the blood tests found

Two trials run for the FDA measured sunscreen absorption. The 2019 pilot gave 4 products to 24 volunteers at 2 mg per square centimetre over 75 percent of the body, 4 times a day for 4 days, with 30 blood samples over 7 days. Every product crossed the agency's threshold of 0.5 nanograms per millilitre of plasma after four applications on day 1, and oxybenzone reached a geometric mean maximum of 209.6 ng/mL.

The follow-up, published in 2020, randomised 48 participants to 4 formulations and tracked 6 filters; 44 finished. All 6 crossed 0.5 ng/mL after a single application at the same dose, with 34 blood samples over 21 days. Oxybenzone reached 258.1 ng/mL from the lotion and 180.1 from the aerosol, homosalate 13.9 to 23.1, octinoxate 5.2 to 7.9, octocrylene 6.6 to 7.8, avobenzone 3.3 to 7.1 and octisalate 4.6 to 5.8.

What the threshold means matters more than the figures. 0.5 ng/mL is the steady state level below which the FDA expects a systemic carcinogenicity study can be waived, set to approximate the level at which an unknown compound would carry a carcinogenic risk below 1 in 100,000 after one dose. Crossing it calls for more toxicology; it is not a finding of harm. No filter was withdrawn: the 12 American ingredients stayed on sale in the data gap category, and the authors wrote that their results do not indicate that people should stop using sunscreen. Covering 75 percent of the body 4 times a day is also heavier than ordinary use, and rash was the commonest adverse event, in 14 of the 48 participants.

White cast, reefs and how to choose

Past the band gap wavelength the oxides reflect up to 60 percent of visible light, so a formula with enough zinc oxide to matter will show on skin, most of all on deeper skin tones. Finer particles scatter less, which is why nano grades exist: the EU permits titanium dioxide (nano) at a median particle size of 30 nm or more and zinc oxide (nano) above 30 nm, neither of them where it could be inhaled, which keeps both out of spray cans.

The reef question has a more careful answer than either camp offers. A 2022 review by the National Academies covered the 17 filters on the American market, 15 organic and 2 inorganic, and put the highest measured water concentrations for five organic filters between 1 and 10 micrograms per litre, most measurements below 1 microgram per litre. Risk characterisation was outside the committee's brief, so it neither convicted sunscreen nor cleared it; it asked the Environmental Protection Agency for a full ecological risk assessment, starting with shallow reefs under heavy recreational use, and named rising temperature as a major stressor in its own right. Laboratory harm shows up mostly below 1,000 micrograms per litre, a hundred times the field measurements, and it shows up for zinc oxide and for titanium dioxide under ultraviolet light too, so swapping to the oxides is not the clean answer it is taken for.

The same report weighed the other side: consistent use of an SPF 30 broad spectrum product reduces melanoma, squamous cell cancer, sunburn and photoageing with high certainty, yet only about a third of the American population uses sunscreen regularly. So take the filter system you will reapply. If you have had a photoallergic reaction, eczema or rosacea, the two metal oxides are the calmer start, and a 3 to 5 day patch test on the inner forearm is worth the wait; see a doctor about a rash that returns each time a product goes back on. Quantity beats chemistry, covered in the dose guide, and the SPF Day Planner handles reapplication.

Reading a sunscreen label properly

Frequently asked

Do mineral filters really block UV by reflecting it?

No. Integrating sphere measurements put the average reflection of zinc oxide and titanium dioxide across the UV range at 4 to 5 percent, which is worth less than an SPF of 2. The other 95 percent or so of their effect comes from absorbing the photons through semiconductor band gap absorbance, the same end result as an organic filter.

Are chemical filters absorbed into the body?

Yes. In a trial of 48 participants, all 6 organic filters tested crossed the FDA threshold of 0.5 ng/mL in plasma after a single application, and oxybenzone reached 258.1 ng/mL. That threshold sets the point at which extra toxicology studies are expected rather than marking a level known to cause harm, and no filter was withdrawn on the strength of it.

Why do European and American sunscreens feel so different?

The EU allows about 30 filter substances, including photostable broad spectrum molecules at up to 15 percent, while the American monograph holds 16 ingredients and has gained none since the late 1990s, because eligibility for it ran through a review of what was already on sale in 1972. Formulators working in Europe have lighter textures available for the same UVA cover.

How can I tell whether a sunscreen covers UVA?

In Europe, UVA printed inside a circle is the industry mark for the criteria in a 2006 Commission recommendation: a UVA protection factor of at least a third of the labelled SPF and a critical wavelength of at least 370 nm, with the UVA factor measured in vitro by the method in ISO 24443. In the United States the phrase broad spectrum means the product reached a critical wavelength of 370 nm or more.

Will a mineral sunscreen always leave a white cast?

Some cast is unavoidable, because past their band gap wavelength the two oxides reflect up to 60 percent of visible light. Nano grades, which the EU allows at a median particle size of 30 nm or more, scatter less visible light, and a tinted formula covers the rest.

Should I avoid certain filters because of coral reefs?

The 2022 National Academies review of 17 UV filters found peak water concentrations of 1 to 10 micrograms per litre, most measurements below 1 microgram per litre, and stopped short of a risk verdict, asking the Environmental Protection Agency for a full ecological risk assessment instead. The same report warned that cutting sunscreen use has its own cost, since consistent SPF 30 use lowers melanoma and squamous cell cancer risk with high certainty.

Sources

• Regulation (EC) No 1223/2009 on cosmetic products, consolidated text of 1 January 2025, Annex VI: list of UV filters allowed in cosmetic products · The EU list itself, read in the consolidated version. Source of the count of about 30 permitted substances across entries numbered to 34, the four separate nanomaterial entries, and every EU maximum quoted here: 25 percent for titanium dioxide and zinc oxide, 5 percent avobenzone, 5 percent ethylhexyl salicylate, 10 percent ethylhexyl methoxycinnamate, 10 percent octocrylene with 9 percent in propellant sprays, 15 percent drometrizole trisiloxane, 10 percent bemotrizinol and bisoctrizole, 10 percent diethylamino hydroxybenzoyl hexyl benzoate, 3 percent for the lowest entry on the list, 6 percent, 2.2 percent and 0.5 percent for benzophenone-3 and 7.34 percent for homosalate in face products with the 1 January and 1 July 2025 deadlines. Also the source of the nano particle size conditions (30 nm median for titanium dioxide and zinc oxide) and the ban on nano grades where the lungs could be exposed. The two amendments cited are Regulation (EU) 2022/1176 of 7 July 2022 and Regulation (EU) 2022/2195 of 10 November 2022, both listed in the amendment table of the same document. The UVA criteria quoted in the text come from a separate instrument, Commission Recommendation 2006/647/EC, which is named there and is not part of this Annex.
• FDA, Sunscreen Drug Products for Over-the-Counter Human Use, proposed rule, Federal Register, 26 February 2019 · The American proposal, read in full text. Source of the 16 active ingredients in the stayed 1999 monograph with their maximum concentrations (avobenzone 3 percent, homosalate 15 percent, octinoxate 7.5 percent, oxybenzone 6 percent, octocrylene 10 percent, titanium dioxide and zinc oxide 25 percent), the proposal that zinc oxide and titanium dioxide are safe and effective at up to 25 percent, that aminobenzoic acid and trolamine salicylate are not, and that 12 ingredients need more data. Also the 0.5 ng/mL threshold and its basis (a carcinogenic risk below 1 in 100,000 after a single dose), the 290 to 400 nm range and the 290 to 320 nm UVB band, the 370 nm critical wavelength requirement for broad spectrum labelling and the proposed UVA I to total UV ratio of 0.7, the proposed SPF 60+ label cap with formulation to SPF 80, the original pre-1972 eligibility condition for the monograph review, and the finding that avobenzone loses between 50 and 90 percent of its effect after 60 minutes of sunlight.
• Cole C, Shyr T, Ou-Yang H. Metal oxide sunscreens protect skin by absorption, not by reflection or scattering. Photodermatol Photoimmunol Photomed. 2016;32(1):5 to 10 · Abstract read through the Europe PMC record for this paper. Source of the integrating sphere measurement showing average reflection of zinc oxide and titanium dioxide across the UV range of only 4 to 5 percent, described in the paper as less than SPF 2, of the conclusion that the remaining protection is semiconductor band gap absorbance, and of the finding that above the band gap, in long UVA and visible wavelengths, the oxides are predominantly reflectors with up to 60 percent reflection.
• Matta MK, Florian J, Zusterzeel R, et al. Effect of Sunscreen Application on Plasma Concentration of Sunscreen Active Ingredients: A Randomized Clinical Trial. JAMA. 2020;323(3):256 to 267 · Abstract read through the Europe PMC record. Source of the 48 participants randomised to 4 formulations with 44 completing, the 6 filters tracked, the application of 2 mg per square centimetre to 75 percent of body surface once on day 1 and 4 times daily on days 2 to 4, the 34 blood samples over 21 days, the crossing of 0.5 ng/mL by all 6 filters after a single application, the maximum plasma concentrations quoted (oxybenzone 258.1 and 180.1, homosalate 13.9 to 23.1, octinoxate 5.2 to 7.9, octocrylene 6.6 to 7.8, avobenzone 3.3 to 7.1, octisalate 4.6 to 5.8 ng/mL), the rash in 14 participants and the authors' statement that the findings do not indicate that individuals should refrain from sunscreen use. The 2019 pilot quoted in the same section is Matta MK, Zusterzeel R, Pilli NR, et al. JAMA. 2019;321(21):2082 to 2091, whose abstract was read through the Europe PMC record MED/31058986: 24 volunteers, 4 products, 2 mg per square centimetre over 75 percent of body surface 4 times daily for 4 days, 30 blood samples over 7 days, all products above 0.5 ng/mL on day 1 and oxybenzone at a geometric mean maximum of 209.6 ng/mL.
• National Academies of Sciences, Engineering, and Medicine. Review of Fate, Exposure, and Effects of Sunscreens in Aquatic Environments and Implications for Sunscreen Usage and Human Health. 2022, Summary · Summary chapter read in full. Source of the 17 UV filters covered (2 inorganic particulates and 15 organic compounds, being the 16 allowed by the FDA plus ecamsule), the highest measured water concentrations of 1 to 10 micrograms per litre for oxybenzone, octocrylene, homosalate, avobenzone and octinoxate with most measurements below 1 microgram per litre, the acute toxicity observations and estimates under 1,000 micrograms per litre with their stated limits, the statement that risk characterisation was outside the committee's scope, the recommendation that the Environmental Protection Agency carry out an ecological risk assessment starting with shallow near shore coral reefs under heavy recreational use, rising temperature as a major stressor in its own right, and the human health conclusions: SPF 30 broad spectrum use reducing melanoma, squamous cell cancer, sunburn and photoageing with a high level of certainty, about a third of the American population using sunscreen regularly and 70 to 80 percent at the beach.

Read on

Tools that fit

This guide is for general information. It does not replace advice from a dermatologist, physician or qualified practitioner, or an examination.