Heavy metals in mineral baby sunscreens

Abstract

Direct measurements of heavy metals in finished sunscreens marketed for babies come only from grey literature: single purchased units tested once by commercial laboratories, with results reported in the product as sold. Seven baby-marked zinc oxide units have quantified lead and cadmium that pass the tests kept in this synthesis. Across those seven, lead ran from 169.2 to 2,140.30 ppb and cadmium from 122.70 to 645.00 ppb, both detected in 7 of 7. Total arsenic was above its 10 ppb reporting limit in three of the seven (485.3, 135.9 and 14.92 ppb) and total mercury above its 5 ppb limit in two (7.88 and 5.29 ppb). All four are total-element results; no source in this collection speciated arsenic or mercury, and none measured inorganic arsenic, methylmercury or hexavalent chromium in any sunscreen.

Three of the seven are separately published single-unit laboratory reports (lead 169.2, 330.0 and 495.1 ppb); the other four are rows in a consumer-advocacy comparison chart. These are the same measurement class, a single unit tested once, and are reported here without pooling. A retail screening of 35 mostly adult sunscreens, two adult peer-reviewed product studies, eight further chart rows without a published report, two additional single-unit laboratory reports whose values or lot records did not reconcile, and applicant impurity data for cosmetic titanium dioxide grades are all discussed in the body, on their own bases, and none is placed in the abstract or a summary table.

No source in this collection measured lead, cadmium, arsenic or mercury crossing skin from a sunscreen, and no penetration study used infant skin. The zinc and titanium penetration data that exist are for adult human skin, in vivo and excised, and for pig skin, and they are consistent across fourteen studies: the mineral particles stay in the stratum corneum, including on barrier-impaired, tape-stripped, occluded and sunburned skin, while a small ionic fraction of zinc reaches the viable epidermis and about 0.01 percent of an applied zinc dose reaches the body. That record is about the filters, not the contaminants, and it is the reason the contaminant question stays open rather than an answer to it. Three surveys of sunscreens not identified as baby or mineral-only report impurity metals and are carried as context; one of them could not have detected the lead concentrations found in the baby units, because its detection limit sits above them. Regulatory numbers that were opened are topical natural-health-product limits, cosmetic-impurity guidance, orientation values and colour-additive raw-material specifications; none is a finished-product heavy-metal limit for sunscreens in the United States, and none was written as a score for a single grey-literature unit. This synthesis states what those documents support. It does not rank products now on the shelf, and it does not set a certification limit.

Product anatomy

A mineral sunscreen is a leave-on emulsion whose sun-protection actives are inorganic particles rather than organic UV filters. The two mineral filters are zinc oxide and titanium dioxide. In the regulatory texts opened here each is permitted at up to 25 percent of the finished product: in the United States as a proposed generally-recognised-as-safe-and-effective active ²⁹, in Canada where a sunscreen containing only these two actives is licensed as a natural health product ³⁰, and in the European Union under Annex VI of the cosmetics regulation (EU Regulation 1223/2009 ³¹). A baby or infant sunscreen is a mineral sunscreen positioned for that age group; none of the regulatory texts sets a distinct heavy-metal specification for the baby version, and the chart and laboratory reports identify products only as baby-marked, not by a separate legal class.

The heavy metals of concern here are not deliberate ingredients. They enter as impurities of the mineral filters and, potentially, of other mineral or earth-derived components. Zinc oxide and titanium dioxide are refined from ores. The titanium dioxide applicant data summarised by ¹² state that trace metals derived from ilmenite, rutile or slag ores are embedded in the lattice of the pigment, which is a claim about where the metal sits, not a measurement of how much reaches a finished sunscreen. That framing matters for this category the way the polyester-catalyst account matters for baby wipes: it names a plausible route by which a metal enters a product, and it is a hypothesis about any given finished unit until the metal is measured in that unit.

The single strongest signal that the filter is the route is a raw-material observation rather than a finished-product one. In the comparison chart a single-ingredient zinc oxide powder sold for making sunscreen at home carried 9,474.70 ppb lead and 506.20 ppb cadmium (¹, row 36), more lead than any finished sunscreen on the same chart. That is one unit of one powder, and it is a raw material, not a sunscreen. It is consistent with the hypothesis that the zinc oxide fraction carries the load, and it does not by itself establish the fraction that ends up in a formulated product.

Measured evidence

Every finished-sunscreen dataset opened for this synthesis is kept on its own basis and its own analytical footing. They are not pooled. The finished baby-product values are single purchased units tested once by commercial laboratories, reported as parts per billion in the product as sold. The retail screening reports milligrams per kilogram as sold from a food-validated digestion. The two peer-reviewed product studies are adult products, one reported in micrograms per kilogram and one in micrograms per gram. The raw-material data are titanium dioxide grades reported as acid-extractable milligrams per kilogram, which is not total content. A value read from one of these is not comparable to a value read from another unless the two share a basis, and where they do not, this paper says so instead of converting across them.

The seven-unit summary below is restricted to the finished baby-marked zinc oxide units whose lead and cadmium results carry no unresolved reconciliation flag and are each counted once. It is the only product summary table in this synthesis. Two further single-unit laboratory reports and the chart rows without a published report are discussed in the text and are deliberately not in this table.

Baby-marked mineral sunscreens, single units

Seven baby-marked zinc oxide units have quantified lead and cadmium as sold, in parts per billion. Three are separately published single-unit laboratory reports; four are chart rows. Non-detects are printed against the laboratory’s reporting limit, 5 ppb for lead, cadmium and mercury and 10 ppb for arsenic, and are not read as zero.

Lead was quantified in 7 of 7 units, from 169.2 to 2,140.30 ppb. Cadmium was quantified in 7 of 7, from 122.70 to 645.00 ppb. Total arsenic was above its 10 ppb reporting limit in 3 of 7 (485.3, 135.9 and 14.92 ppb) and below it in the other four. Total mercury was above its 5 ppb reporting limit in 2 of 7 (7.88 and 5.29 ppb) and below it in the other five. Sources for this table are the comparison chart ¹ and the three laboratory reports ^{2, 3, 4}. The three laboratory-report units are the same tests as three of the chart’s baby-marked rows and are counted once here, as their own reports rather than as chart rows.

The rest of the measured record

The comparison chart holds more than the seven summarised units. Its header counts, which reconcile with its rows 1 to 35, are lead positive in 33 of 35 sunscreens, cadmium in 29 of 35, arsenic in 13 of 35 and mercury in 5 of 35 ¹. Across the twenty-seven rows charted as zinc oxide sunscreens, lead ran from 118.00 to 4,612.50 ppb and cadmium from 9.67 to 645.00 ppb. The two products charted as chemical-filter sunscreens were non-detect for all four metals. Eight rows are marked to be published and have no laboratory report in this collection; none of them is a baby-marked row, and they are not used as range ends here. The chart is a single unit per product, community-nominated, one laboratory, and a draft, with no method validation reported for zinc oxide matrices.

Two further single-unit laboratory reports name baby zinc oxide products but are held out of the summary table. One reports lead 1017.5 and cadmium 356.1 ppb; its internal lot number is printed out of sequence with the others in the same series, an unresolved records anomaly. The other reports lead 1251.1 and cadmium 575.5 ppb; it carries the same product-line name as a unit in the retail screening described below, and the two disagree once put on the same basis, so this synthesis keeps them apart and does not treat either as confirming the other. Both units sit within the same order of magnitude as the seven summarised units and neither changes the qualitative finding, but neither is placed in the abstract, the range ends or a table.

A 2026 company screening bought 35 retail sunscreens, mostly adult, and sent one unit of each to an accredited laboratory ⁵. Lead was quantified in all 35, from 0.006 to 4.960 mg/kg; cadmium in 21, from 0.005 to 0.603 mg/kg; total arsenic in 9, from 0.005 to 0.189 mg/kg; and total mercury in 3, from 0.004 to 0.009 mg/kg. Every unit with quantified cadmium had lead of at least 0.125 mg/kg, and every cadmium-non-detect unit had lead of at most 0.079 mg/kg; the report reads that co-occurrence as a pointer to a shared ingredient while stating that the panel cannot identify it. The four units whose certificate names include the word mineral were all in the cadmium-positive group. This screening used a digestion validated for foods, whose completeness for mineral particles the report states was not evaluated, and it notes that any shortfall would understate content. Its milligram-per-kilogram results are on a different digestion basis from the single-unit baby reports and are not pooled with them or placed in the summary table.

A Korean consumer federation tested 17 infant and child sunscreens against the national cosmetic standard and reported every one compliant or not detected for six metals, without printing any concentration, detection limit or limit value ⁶. Those results support a statement of compliance in that programme; they are not concentrations and cannot be placed beside the numeric results above.

Three further peer-reviewed surveys measured impurity metals in finished sunscreens that are not identified as baby products and not identified as mineral-only. They are reported here as context. None of them is a baby mineral sunscreen, none of them states whether its products used mineral or organic filters except where noted, and none of their values belongs in the seven-unit summary table or in any range on this page.

Zmozinski and colleagues measured total lead and total chromium in eleven coded commercial sunscreens from Porto Alegre, Brazil, by high-resolution continuum source graphite furnace atomic absorption spectrometry with direct solid sampling ⁷. Total chromium ran from 0.10 ± 0.01 to 0.65 ± 0.07 µg/g across the eleven. Total lead ran from 0.027 ± 0.002 to 1.75 ± 0.10 µg/g, with two samples printed as below the limit of quantification. Detection limits were 3 µg/kg for lead and 1 µg/kg for chromium, and quantification limits 9 and 4 µg/kg, calculated for a 4.5 milligram sample; a subset re-run after digestion agreed with the direct results at 95 percent confidence. This is the most methodologically complete lead measurement of finished sunscreens in the collection, and its lead range overlaps the baby-unit range almost exactly, from roughly 0.03 to 1.75 µg/g against 0.169 to 2.14 µg/g. That overlap is worth naming and is not a confirmation: the products are a different market, are not identified as mineral, and are not baby products, so the agreement in magnitude is a reason to run the measurement properly on baby mineral products rather than evidence about them.

Zachariadis measured ten elements in five coded commercial sunscreens from Greece by inductively coupled plasma atomic emission spectrometry after nitric, hydrochloric and hydrofluoric acid digestion ⁸. Titanium, as the element, reached 8.13 x 10^3 and 18.5 x 10^3 µg/g in the two products that listed titanium dioxide and was below 0.2 µg/g in two others, which is the expected filter signature. Lead was below the detection limit in all five. That last result carries no weight here, and the reason is arithmetic rather than interpretive: the lead detection limit was 3.2 µg/g, which is above the highest lead concentration measured in any baby unit in this file, 2.14 µg/g. A method that cannot see 2 µg/g of lead cannot report on the concentrations this page is about, so its non-detects are not an upper bound for anything discussed here. Chromium, copper, manganese, iron, zinc, aluminium, magnesium and boron varied widely among the codes, including high iron, zinc, copper, manganese and chromium in one code, and those per-code values are not transcribed on the source page and so are not given here.

Ali and colleagues measured eight elements in fifteen coded sunblock products bought in Lahore, Pakistan, each digested in triplicate and read by flame atomic absorption, with products described as above cited standards re-measured by inductively coupled plasma optical emission spectrometry ⁹. Chromium and arsenic were reported as not detected in all fifteen, with no numeric non-detect limit printed, so neither absence is an upper bound. Among the cells that are printed with their product code, Table 1 gives lead of 5.761 ± 0.0901 mg/kg in code SB8, nickel of 5.2 ± 0.3773 mg/kg in code SB15 and titanium of 4.1 ± 0.3295 mg/kg in code SB9. This study is held to context for a stronger reason than its market: its own numbers do not agree with each other. The abstract, the flame atomic absorption narrative and Table 1 print different values for the same cells, differing in the third decimal for lead in SB8 and in the first for nickel in SB15 and titanium in SB9, the full zinc and aluminium series are not printed, and Section 2.6 and Sections 3.2.4 and 3.2.5 disagree over whether a 20 mg/kg zinc and aluminium figure belongs to the United States Food and Drug Administration or to the European Union. Limits the paper attributes to a 1976 European directive and to that agency are the authors’ citations and are not entered as instruments anywhere in this synthesis. No range is drawn from this study and no maximum is claimed from it, because the disagreements mean the printed extremes cannot be relied on.

Two peer-reviewed studies measured metals in finished sunscreens, both adult products. Rodríguez-Romero and colleagues characterised one SPF 50 spray containing nano titanium dioxide and reported lead 743, cadmium 16.7 and nickel 47.5 µg/kg as sold, alongside percent-level titanium (1.48 x 10^7 µg/kg) and aluminium ¹⁰; arsenic and mercury were not measured, and the paper prints no standard deviation for the product row despite stating triplicate analysis. Tovar-Sánchez and colleagues found titanium (6,302 to 14,855 µg/g) in three of seven sunscreens and zinc (72,688 µg/g) in one, and reported lead, cadmium, chromium, nickel and ten other elements as not detected, without printing a detection limit ¹¹. Because no detection limit is given, that absence is a method statement rather than an upper bound, and it cannot be read as evidence those products were free of lead or cadmium. Both studies are adult products and bear on baby mineral sunscreens only as peer-reviewed context that lead and cadmium occur at measurable levels in at least one nano titanium dioxide sunscreen, and that titanium and zinc dominate the mass where they are present.

What was measured, by metal

The table records where each panel metal was measured in a finished baby-marked mineral sunscreen and what was not measured. Values not in the seven-unit summary are named here on their own basis. Concentrations are in the product as sold unless stated.

Toxicology

Toxicology here is the hazard each metal is known for, the endpoint that usually drives that hazard, and what, if anything, is known about the metal on infant skin or in a sunscreen. A classification or a reference dose derived from ingestion or inhalation is not a dermal sunscreen dose. Where a dermal number was not in a document opened for this paper, the cell says so.

Infant skin is the exposure surface, and this paper does not assign it an absorbed fraction. The reasons the infant surface matters are set out in a narrative review opened for this collection ²⁸: the review reports that the infant stratum corneum is about 30 percent thinner than the adult and the viable epidermis about 20 percent thinner, that barrier properties come to resemble adult skin about two years after birth, and that an infant’s surface area per body weight is 2.3 times the adult value, so a product applied over a large body surface reaches proportionally more skin. It also notes that sunscreens are often applied over a large body surface area and that the American Academy of Pediatrics restricts infant sunscreen use to areas not covered by clothes to limit absorption of sunscreen ingredients. These are reasons to measure the dermal route for metals in a sunscreen. They are not a measured fraction, and the review reports no heavy-metal absorption value.

Lead

Lead is a developmental neurotoxicant. The endpoint that dominates infant and child assessments is harm to the developing nervous system, with no opened threshold in this file below which a dermal dose would be called without effect. The sunscreen evidence is a concentration, not that dose. Lead was quantified in all seven summarised baby units, from 169.2 to 2,140.30 ppb as sold, and in a raw zinc oxide powder at 9,474.70 ppb. No source measured blood lead, dermal absorption or hand-to-mouth transfer from a sunscreen. Calling any of these a violation of a cosmetic or natural-health-product limit would require the same basis as the instrument and a sample taken while the instrument applied; the finished-product limits opened here are on a total-content or acid-extractable basis and were not written for a single nominated unit.

Arsenic

The arsenic hazard that regulatory food work isolates is inorganic arsenic, with carcinogenicity and, at lower exposures, developmental and dermal effects attaching to the inorganic species rather than automatically to every total-arsenic number. Every arsenic result in this collection is total arsenic. Three of the seven summarised units were above the 10 ppb reporting limit (485.3, 135.9 and 14.92 ppb). Inorganic arsenic in a sunscreen: no data found. A total-arsenic detect is a reason to speciate. It is not an inorganic-arsenic result, and it is not compared here with any inorganic-arsenic benchmark.

Mercury

Methylmercury is the organic species associated with neurodevelopmental harm after ingestion, especially from fish; elemental and inorganic mercury have different kinetics, including dermal and inhalation routes for some forms. Every mercury result here is total mercury. Two of the seven summarised units were above the 5 ppb reporting limit (7.88 and 5.29 ppb); the other five were below it. Methylmercury in a sunscreen: no data found. A total-mercury result below the reporting limit does not prove the absence of a trace beneath that limit, and a total-mercury detect is not a methylmercury result.

Cadmium

Cadmium’s critical long-term endpoint in dietary assessments is kidney tubular dysfunction, with bone effects at related exposures, and that record is an oral, cumulative one. Cadmium was quantified in all seven summarised units, from 122.70 to 645.00 ppb as sold, and in the raw zinc oxide powder at 506.20 ppb. No dermal-absorption fraction for cadmium from a sunscreen was opened. The oral kidney endpoint is not a dermal sunscreen dose, and this paper does not convert one into the other.

Chromium

Chromium toxicity depends on oxidation state. Trivalent chromium is poorly absorbed by mouth; hexavalent chromium is the sensitiser and the carcinogenic species in the occupational record, which is largely inhalation and, for sensitisation, skin. No finished baby sunscreen in this collection was measured for chromium. One adult peer-reviewed study lists chromium among elements reported not detected, without a detection limit, so that absence is not an upper bound. Hexavalent chromium in a sunscreen: no data found. A total-chromium number, had one been measured, could not be read as hexavalent chromium.

Nickel

Nickel is the most common cause of metal allergic contact dermatitis. Sensitisation is acquired through skin exposure, and later contact can elicit a rash; the clinical literature expresses both sensitisation and elicitation thresholds as release per area over time, not as a bulk concentration. No baby mineral sunscreen in this collection was measured for nickel. One adult spray reported 47.5 µg/kg as sold, and one adult study reported nickel not detected without a limit. No opened study shows nickel in a sunscreen crossing skin or eliciting dermatitis, and no bulk concentration here has been converted to a release per square centimetre.

Tin

Inorganic tin in food is an acute gastric irritant at intakes far above ordinary trace levels, a record about canned food rather than skin; organotin compounds are a separate hazard class used as biocides, stabilisers and some catalysts. No tin measurement in any sunscreen, inorganic or organic, was found in this collection. No dermal threshold for tin from a sunscreen was opened.

Aluminium

Aluminium neurotoxicity and bone toxicity are documented at high systemic exposures, particularly when the kidney cannot clear the metal, as in infants on parenteral nutrition. That record does not transfer to a dermal sunscreen exposure by analogy. No baby mineral sunscreen in this collection was measured for aluminium. The adult spray reported percent-level aluminium (1.67 x 10^6 µg/kg as sold) as a formulation-level constituent, and one adult study reported 342 and 1,071 µg/g in two products; neither is a baby product and neither is a dermal-absorption measurement. Aluminium crossing skin from a sunscreen: no data found.

The mineral filters: zinc and titanium penetration

Zinc oxide and titanium dioxide are the sun-protection actives, not contaminant metals, but the penetration studies opened for this collection are all about these two, and they bound how far a topically applied mineral particle travels. They are relevant to this synthesis because the contaminant metals of concern ride in or on these same particles, and because the studies show what dermal work on this product class has and has not established. The studies are grouped below by the skin they used, because that is the distinction a reader has to keep: adult human skin in a living body, adult human skin removed from a body, animal skin, and deliberately damaged skin are four different claims. None of the fourteen used infant skin. None measured lead, cadmium, arsenic or mercury crossing skin.

Adult human skin, in vivo

Two isotope-tracer studies dosed living adults and measured what reached the body. In an outdoor trial, twenty adults applied sunscreens made with zinc oxide enriched above 99 percent in the stable isotope 68Zn, one formulation at about 19 nanometres and one at about 110 nanometres, over five outdoor days at application rates of 3.7 to 5.3 milligrams per square centimetre, with blood and urine taken before, during and six days after ¹³. Small increases in tracer zinc appeared in blood and urine, the largest blood signal was in the women using the nanoparticle formulation, and the authors state that the chemical form crossing the skin, particle or dissolved ion, was not determined. A later pilot applied about 30 nanometre zinc oxide enriched to 52 percent 68Zn twice daily for five days in winter without ultraviolet exposure and described absorption as minimal, at 0.01 percent of the applied dose ¹⁴. That figure is the only systemic absorbed fraction for a mineral filter in this collection, and it is read with one caveat: some readings of that document show a less-than glyph before the number, and the token verified during ingest is 0.01 percent of the applied dose. A conference report from the same programme found tracer zinc in blood after the application period at about one thousandth of total blood zinc, still rising afterwards ¹⁵. Separately, repeated application of zinc oxide nanoparticles in five adult volunteers left the viable-epidermis particle signal indistinguishable from untreated and vehicle controls ¹⁶.

Titanium behaves the same way in living adults. Adults given three titanium dioxide sunscreens, one of which also contained zinc oxide, at about 0.5 to 1.0 milligrams per square centimetre for two hours or at the 2 milligram-per-square-centimetre sun-protection-factor test thickness, had high titanium and, where zinc oxide was present, high zinc in the stratum corneum, falling to undetectable titanium and physiological zinc below it ¹⁷. Occluding the site for 48 hours and tape stripping it did not produce a printed titanium signal in the subcorneal epidermis.

Adult human skin, excised or reconstructed

On excised adult human skin, zinc oxide nanoparticles labelled with a rare zinc isotope stayed in the stratum corneum while dissolved zinc reached the viable epidermis, where exogenous zinc was 1.0 ± 0.3 µg/mL against 4.3 ± 0.7 µg/mL of endogenous zinc, below the concentration cytotoxic to cultured keratinocytes ¹⁸. Zinc oxide at 10 percent by weight, in a triglyceride vehicle and in aqueous suspensions at pH 6 and pH 9, applied to heat-separated human epidermis at 10 milligrams per square centimetre for up to 48 hours, did not penetrate intact viable epidermis as particles, while labile zinc fluorescence increased and was highest with the pH 6 vehicle ¹⁹; that experiment also recorded viable-epidermis pH falling from 6.83 ± 0.42 to 5.10 ± 0.20 under the pH 6 buffer, which are pH readings and not zinc concentrations. Commercial nano zinc oxide sunscreen formulations on human skin left zinc in the viable epidermis that was not identifiable as nano zinc oxide particles, with zinc in the receptor solution at 0.07 µg/mL after 48 hours ²⁰. Beneath a reconstructed human epidermis, zinc reached the medium at about 2,400 to 2,700 µg/L against roughly 25 to 37 µg/L in controls, while no zinc oxide particles were seen in the tissue and titanium was not distinguishable from the controls, from which the authors conclude zinc crosses as ions ²¹. Titanium dioxide nanoparticles in synthetic sweat did not reach the receiving solution of Franz cells; they were found in the epidermis of intact skin at 0.47 ± 0.33 µg/cm2, and needle-abraded skin as a whole held 0.53 ± 0.26 µg/cm2 ²².

Animal skin

Pigs with a moderate ultraviolet-B sunburn were treated with four sunscreens, 10 percent coated titanium dioxide in oil-in-water and in water-in-oil, and 5 percent coated and 5 percent uncoated zinc oxide in oil-in-water ²³. Titanium in the in vitro perfusate stayed below 0.1 µg/mL and below 0.1 percent of the applied dose in every sample tested, and zinc in perfusate matched untreated controls. Pig skin is the closest common animal model to human skin for dermal work and it is still not human skin, and a moderate sunburn is not infant diaper-area or infant body skin. In a separate model, ultraviolet-B damaged mouse skin took up more zinc, direction only, with the numeric values in a supplement not in this collection ²⁴.

Damaged, occluded and impaired skin

Barrier state is the condition under which a reader might expect the particle findings to fail, and in this collection it has now been tested directly. Topical nano zinc oxide at 10 percent weight per volume, applied to volunteers at the standard 2 milligram-per-square-centimetre sun-protection-factor dose for six hours, did not penetrate into the viable epidermis of intact skin or of barrier-impaired skin, with or without occlusion ²⁵. Tape stripping and 48 hours of occlusion likewise produced no subcorneal titanium signal in living adults, and needle abrasion did not move titanium into the receiving solution of a diffusion cell. Against that, zinc within skin rose in impaired compared with intact skin, by about 3.5-fold in a sweat vehicle and about 3-fold in a triglyceride vehicle ²⁰, the ultraviolet-B mouse model took up more zinc, and the pig work used sunburned skin throughout. The pattern that survives is a split: barrier damage raises the amount of ionic zinc held in skin, and it has not been shown to let the particles themselves through. Two reviews supply the size framing rather than new measurements, adopting 4 to 20 nanometres as the range that may cross intact and damaged skin and 21 to 45 nanometres for damaged skin ²⁶, and defining the nanoparticle designation for these filters as below 100 nanometres ²⁷.

What none of this covers

Across fourteen studies, four skin types, two mineral filters, and intact, tape-stripped, abraded, occluded, barrier-impaired and sunburned conditions, the finding is consistent: the mineral particles stay in the stratum corneum, a small ionic zinc fraction reaches the viable epidermis, and roughly 0.01 percent of an applied zinc dose reaches the body. That consistency is worth stating plainly because it cuts against the intuition that a damaged barrier would let the particles through; on the evidence here it does not, though it does raise the ionic zinc held in skin.

None of that is an answer to this page’s question, and the reason is not a matter of degree. Every one of those studies measured the filter, zinc or titanium, which is present at percent levels and is the ingredient the product exists to deliver. The metals this page is about are present at parts per billion, they are impurities rather than actives, and no study in this collection measured any of them crossing skin from any sunscreen. A lead atom embedded in a titanium dioxide lattice and a titanium atom in the same lattice do not have to behave alike once the particle sits in the stratum corneum for hours in the presence of sweat, and nothing opened here tests the difference. Nor did any study use infant skin. The particle record is the closest dermal evidence this collection holds, it is an argument for measuring the contaminant metals directly, and it is not a substitute for having done so.

Raw materials

The contaminant metals enter a mineral sunscreen through its mineral components, and the collection holds more on the raw materials than on the finished baby products. It is kept separate because a raw-material concentration is not a finished-sunscreen concentration, and because the raw-material bases differ from the finished-product bases.

Zinc oxide is represented by one unit: the single-ingredient powder in the comparison chart, at 9,474.70 ppb lead and 506.20 ppb cadmium as sold (¹, row 36). It is one unit of one powder sold for home formulation, and it carried more lead than any finished sunscreen on the same chart. It is consistent with the zinc oxide fraction being a lead and cadmium carrier and does not fix the fraction that reaches a formulated product.

Titanium dioxide is represented by applicant impurity data for 84 cosmetic grades summarised by the European committee ¹². Most grade rows print specification ceilings rather than measured numbers. The committee’s summary maxima for pigmentary grades are, on an HCl-soluble basis, arsenic at most 1, lead at most 10, antimony at most 2, mercury at most 1 and cadmium at most 1 mg/kg; for nano grades the summary gives arsenic below 1, lead below 10, antimony below 2 and mercury below 1 ppm, with cadmium not provided. Three grades print numeric profiles: two pigmentary grades at 0.3 and 0.8 mg/kg lead, both with arsenic, antimony, mercury and cadmium below 0.1, and one nano grade at 3 mg/kg lead, 0.5 antimony, 0.1 mercury and arsenic below 1 mg/kg. Arsenic, lead, antimony and cadmium in these profiles are extracted in boiling dilute hydrochloric acid rather than totally digested, so they are not total content and cannot be compared directly with a total-digestion finished-product result. Mercury in the same profiles is measured after a sulfuric and nitric acid digestion, although the committee’s summary table labels it HCl-soluble. The earlier committee opinions on nano zinc oxide and nano titanium dioxide record that no analytical impurity data were submitted at all, and the nano titanium dioxide opinion notes an applicant claim of conformity with pharmacopoeial and colour-additive ceilings without supporting data.

The picture the raw-material record supports is narrow and consistent with the finished-product signal: the mineral filters are refined from ores that carry trace lead, cadmium, arsenic and antimony; the one zinc oxide unit measured for total content was high in lead and cadmium; and the titanium dioxide record is mostly applicant ceilings on a partial-extraction basis. It does not quantify how much of any metal a given finished baby sunscreen inherits from its filters.

Exposure framing

A sunscreen exposure has at least four inputs: how much metal is in the product, how much of the product is applied to skin, how long it stays there, and what fraction is absorbed. Application amount and reapplication frequency are real variables for a sunscreen used on a child over a day at the beach, and this paper assigns them no numbers, because no primary time-and-motion or diary study of infant sunscreen use was opened, and no dermal-absorption fraction for any of these four metals from a sunscreen was measured.

An illustrative dose is not offered. The rule used here is that every input has to come from a document that was opened, and no opened source supplied, together, an applied mass, a use frequency, a contact time and a dermal-absorption fraction for lead, cadmium, arsenic or mercury from a sunscreen. Inventing the missing inputs would produce something shaped like a dose that is not one. The zinc and titanium penetration studies bound the behaviour of the filter particles on adult skin; they do not supply an absorbed fraction for the contaminant metals, and the one infant-skin document in the collection is a barrier-function review, not an absorption measurement. The gap is the finding.

What can be said without arithmetic is qualitative and still useful. In the single baby units tested for them, lead and cadmium were present in every one, at hundreds to low thousands of parts per billion for lead and roughly a hundred to several hundred for cadmium, in the product as sold. Total arsenic and total mercury were detected in a minority of the same units. The raw zinc oxide powder was higher in lead than any finished sunscreen tested. The dermal step is unmeasured for these metals. Until it is measured, the honest consumer statement is that a concentration in a sunscreen is a concentration in a sunscreen, and that these are single units rather than a market.

Regulatory landscape

A mineral sunscreen sits in more than one legal frame across the texts opened here, and the frames are not interchangeable. In the United States a sunscreen is an over-the-counter drug; in Canada a titanium-dioxide-and-zinc-oxide sunscreen is a natural health product; in the European Union a sunscreen is a cosmetic. The limits below are cited only in the unit and the legal character of the instrument. None of them is a finished-product heavy-metal limit for sunscreens in the United States, and no such limit appears in any source in this collection. The single grey-literature units are not scored against these instruments, because the bases differ and the instruments were not written for a single nominated unit.

Canada applies the topical limits of its Natural Health Products quality guide to sunscreens licensed under the primary sunscreen monograph: on a finished-product basis, 10 ppm lead, 3 ppm arsenic, 3 ppm cadmium, 1 ppm total mercury and 5 ppm antimony ³⁰. The monograph’s youngest subpopulation is infants from 6 months. Health Canada’s separate cosmetic-impurity guidance treats the same lead, cadmium, mercury and antimony figures as technically avoidable levels for cosmetics, with arsenic at 3 ppm.

In the United States, the proposed sunscreen rule would recognise zinc oxide and titanium dioxide as generally safe and effective at up to 25 percent each and proposes no heavy-metal impurity limit for the actives or the finished product ²⁹. The existing labelling it cites reads, for children under 6 months, ask a doctor. The two colour-additive specifications that do carry metal limits apply to titanium dioxide and zinc oxide as colour-additive raw materials, not as sunscreen actives, and on a raw-material basis: titanium dioxide at 10 ppm lead, 1 ppm arsenic, 1 ppm mercury and 2 ppm antimony, with lead, arsenic and antimony measured in an acid extract rather than by total digestion; zinc oxide at 20 ppm lead, 15 ppm cadmium, 3 ppm arsenic and 1 ppm mercury.

In the European Union, the cosmetics regulation prohibits lead, arsenic, cadmium, mercury and antimony and their compounds as intentional ingredients and permits only technically unavoidable traces under Article 17, setting no numeric trace ceiling (EU Regulation 1223/2009 ³¹). Annex VI permits titanium dioxide and zinc oxide at up to 25 percent each and sets purity and total-impurity conditions for the nano forms, titanium dioxide at least 99 percent pure and nano zinc oxide at least 96 percent pure with impurities limited to carbon dioxide and water and any others below 1 percent in total, but no element-specific lead, cadmium, arsenic or mercury limit. Numeric trace ceilings for cosmetics in the EU frame are national orientation values, such as the German figures of 2.0 mg/kg lead, 0.1 mg/kg cadmium, 0.5 mg/kg arsenic, 0.1 mg/kg mercury and 0.5 mg/kg antimony for cosmetics in general ³², which are orientation values rather than statutory limits and were derived from monitoring that did not include sunscreens.

Several benchmarks named by the grey-literature sources are not in this collection and are not used as comparators here. The comparison chart labels lead of 1,000 ppb and above possibly illegal under a Washington State cosmetics act, while the same page reports that state administrators treat sunscreens as outside that limit; the instrument is not in this collection, and no legal comparison is drawn. The retail screening cites a US FDA lead guidance value for cosmetics, further Washington State figures and a California Proposition 65 lead dose level; those instruments are likewise not in this collection and are not placed against any product value.

Implications

For parents

A parent can take three statements from this file, and should not take a fourth. In the single baby mineral sunscreens that have been tested for them, lead and cadmium were found in every one, at concentrations in the product ranging into the low thousands of parts per billion for lead. Total arsenic and total mercury were detected in a minority of those same units, and inorganic arsenic, methylmercury and hexavalent chromium were not measured at all. The one raw zinc oxide powder tested for total content carried more lead than any finished sunscreen on the same chart, which points at the mineral filter as the likely route.

The statement not to take is a risk verdict. No opened study follows a sunscreen from the bottle onto infant skin and into the body, and no study measured how much of any of these metals crosses skin. These are single purchased units, not a survey of the market, and a concentration in a product is not a dose to a child. Sun protection itself is a health measure, and the tradeoff between using a sunscreen and avoiding one is a caregiving decision this paper does not make. What the file supports is that the metal content of these products has been measured too seldom, and on too few units, for anyone to describe the category.

For manufacturers

The mineral filter is where the metal most plausibly enters, and it is where measurement should start. The single zinc oxide powder result and the co-occurrence of lead and cadmium in the retail screening both point at the zinc oxide fraction, and the titanium dioxide record is mostly applicant ceilings on a partial-extraction basis rather than measured total content. A finished-product number cannot be assigned to the zinc oxide, the titanium dioxide or another mineral component without measuring the raw materials and the formulated product on the same basis.

Basis has to be recorded and held constant. The single-unit reports are parts per billion by an unstated digestion, the retail screening is a food-validated digestion whose recovery from mineral particles was not checked, the colour-additive limits are acid-extractable, and the titanium dioxide applicant data are HCl-soluble. A completeness check on mineral-particle digestion is the difference between a number that can be compared with a total-content limit and one that cannot.

For regulators

None of the instruments opened here is a finished-product heavy-metal limit for sunscreens in the United States, and the ones that exist elsewhere are not interchangeable. Canada reaches sunscreens through a topical natural-health-product limit; the EU reaches them through a general cosmetic trace-impurity clause with national orientation values and nano purity conditions but no element-specific metal limit; the US colour-additive figures are raw-material specifications for a different use of the same minerals. A sunscreen standard that copied only one of these would not match the others, and a raw-material acid-extractable ceiling is not a finished-product total-content limit.

Speciation rules that already exist for food should not be relaxed because the matrix is a sunscreen. Total arsenic is not inorganic arsenic, and total mercury is not methylmercury. Every arsenic and mercury result in this collection is a total, and no source speciated either in a sunscreen. Where a government has no number for a metal in this product, the gap should stay a gap rather than being filled by borrowing a limit written for another matrix or another use.

For researchers

The next useful study is small and specific. Measure the finished sunscreen and its mineral raw materials on the same total-digestion basis, with a completeness check for mineral particles, and report the result in milligrams per kilogram as sold rather than only as an instrument concentration. Include lead, cadmium, arsenic and mercury, speciate arsenic and mercury, and add chromium, nickel, tin and aluminium so the panel is not silent on the metals no one has yet measured in a baby product. Sample more than one unit and more than one lot, and draw the sample from a defined frame rather than from community nomination. Then, and only then, measure the dermal step, on skin models that include the compromised and occluded conditions relevant to infants, or state explicitly that the absorbed fraction is unknown.

Two design requirements follow and should be stated as requirements rather than preferences. A quantification limit at or below roughly 0.01 milligrams per kilogram is necessary for the lead question, because at least one published sunscreen survey reported no lead using a detection limit above the highest concentration found in any baby unit here, and a study repeating that error would produce a reassuring null that means nothing. And the dermal experiment has to trace the impurity, not the filter. Fourteen studies now show where zinc and titanium go, using isotope tracers, electron microscopy and elemental mapping, precisely because those elements are abundant enough to follow. Lead at a part per billion in a matrix containing percent-level zinc is a harder measurement, and it is the one that has never been attempted. An isotopically labelled lead spike in a formulated mineral sunscreen, on excised human skin under occlusion, would answer in one experiment what the entire filter-penetration literature cannot.

Until that study exists, synthesis has to keep repeating the basis problem and the single-unit problem. The measurements that exist are real, and none of them is a market, a dose, or a total-content comparison with a finished-product limit. That is enough to justify the work and not enough to finish it.

Data gaps and research agenda

No peer-reviewed measurement of heavy metals in a finished sunscreen marketed for babies is in this collection; the only baby-product values are single commercially tested units. The added surveys sharpen that gap rather than closing it: three peer-reviewed impurity surveys of finished sunscreens were added, and not one identifies its products as baby or as mineral-only, so the peer-reviewed record still contains no baby mineral sunscreen. No source measured inorganic arsenic, methylmercury or hexavalent chromium in any sunscreen or raw material. No source measured lead, cadmium, arsenic or mercury crossing skin from a sunscreen, and no penetration study used infant skin; the zinc and titanium penetration data are for adult human skin, excised human skin and pig or mouse skin.

Two gaps are newly visible because the new sources exposed them. The first is a detection-limit gap: a survey reporting no lead in five sunscreens used a limit of 3.2 µg/g, above the highest baby-unit lead value in this file, so parts of the existing peer-reviewed literature are incapable of detecting the concentrations at issue and their negative results should not be read as reassurance. Any future survey needs a quantification limit at or below roughly 0.01 µg/g to speak to this range. The second is a filter-versus-impurity gap: fourteen studies now measure how the mineral filters behave on skin and none measures how an impurity metal carried by those filters behaves, which is a different question about a different element at a concentration four to six orders of magnitude lower, since the filters sit at percent levels and the measured lead sits below 2.2 mg/kg. Chromium, nickel, tin and aluminium were not measured in any baby mineral sunscreen. Digestion differs across the sources that do exist, from unspecified commercial methods to a food-validated method to acid extraction, so their values are not directly comparable, and none is shown to recover metal locked in mineral particles. The titanium dioxide raw-material data are applicant-supplied and mostly ceilings on a partial-extraction basis; the zinc oxide raw-material data are one unit from grey literature and a colour-additive specification. The Korean compliance results carry no concentrations.

The research agenda follows those gaps, in order. First, total-digestion content in milligrams per kilogram for finished baby mineral sunscreens and their zinc oxide and titanium dioxide raw materials on the same basis, with a mineral-particle completeness check, across lead, cadmium, arsenic and mercury and adding chromium, nickel, tin and aluminium. Second, speciation of arsenic and mercury, and of chromium if it is measured. Third, a sampling frame that is not community nomination, with more than one unit and lot per product. Fourth, a dermal-absorption experiment for the contaminant metals on skin models relevant to infants, or an explicit statement that the fraction is unknown. Fifth, the mineral raw materials characterised apart from the finished product, so the route the metals take can be attributed rather than assumed.

Peer review state

This synthesis claim has not yet been evaluated by external reviewers. Verdicts will be added here as named domain experts, listed at Curators and conflict-of-interest disclosure, complete their review. Reviewer verdicts are recorded on this page.

The Heavy Metal Index publishes synthesis claims as preprints — before external review completes — with the review state visibly tracked. Until at least one external verdict is recorded below, a synthesis page is a preprint, not a peer-reviewed work. External review accumulates over time, and the credibility of the claim is partly the cumulative result of that visible review.

How to cite

Pendergrass, K. (2026). Heavy metals in mineral baby sunscreens. Institute of Contaminant Standards. https://doi.org/10.83376/hmi.heavy-metals-in-mineral-baby-sunscreens

References

1. Chart comparing the toxicant profiles of popular sunscreen productsRubin, T., & Lead Safe Mama, LLC · tamararubin.com (Lead Safe Mama, LLC), posted 28 May 2026; chart Draft 2.1 updated 19 April 2026 · 2026 · tamararubin.comNGO report
2. Laboratory test of a baby mineral sunscreen lotion with prebiotics, SPF 55 (laboratory case 2507165)Rubin, T., & Lead Safe Mama, LLC · tamararubin.com (Lead Safe Mama, LLC), posted 2 August 2025, updated 28 August 2025 · 2025 · tamararubin.comNGO report
3. Laboratory test of a sensitive-skin baby mineral sunscreen lotion, SPF 50 (laboratory case 2505470)Rubin, T., & Lead Safe Mama, LLC · tamararubin.com (Lead Safe Mama, LLC), posted 12 June 2025, updated 16 August 2025 · 2025 · tamararubin.comNGO report
4. Laboratory test of a baby mineral sunscreen lotion, SPF 50 (laboratory case 2506212)Rubin, T., & Lead Safe Mama, LLC · tamararubin.com (Lead Safe Mama, LLC), posted 14 June 2025, updated 19 June 2025 · 2025 · tamararubin.comNGO report
5. 35 sunscreens tested: residual solvents and heavy metalsOasis Health · Oasis Science Testing Program, oasishealth.app, findings report, Order No. 594282, last updated September 2026 · 2026 · www.oasishealth.appIndustry report
6. Test and evaluation results for infant and child sunscreensDaejeon-Chungnam Consumer Federation · Consumer24 (Korea Fair Trade Commission portal), posted 12 April 2021 · 2021 · www.consumer.go.krNGO report
7. Determination of lead and chromium in sunscreens by high-resolution continuum source graphite furnace atomic absorption spectrometry with direct solid samplingZmozinski, A. V., Passos, L. D., Damin, I. C. F., & Vale, M. G. R. · 2016Peer-reviewed
8. Multi-element method for determination of trace elements in sunscreens by ICP-AESZachariadis, G. A., & Sahanidou, E. · 2009Peer-reviewed
9. Determination of heavy metals and their associated health risk assessment in sunblock productsAli, M., Shahzad, L., Hayyat, M. U., Sharif, F., Ghafoor, G. Z., & Nasir, R. · products purchased in Lahore, Pakistan · 2023Peer-reviewed
10. Sunscreens as a new source of metals and nutrients to coastal watersRodríguez-Romero, A., Ruiz-Gutiérrez, G., Viguri, J. R., & Tovar-Sánchez, A. · Environmental Science & Technology 53(17):10177-10187 · 2019 · doi.org/10.1021/acs.est.9b02739Peer-reviewed
11. Sunscreen products as emerging pollutants to coastal watersTovar-Sánchez, A., Sánchez-Quiles, D., Basterretxea, G., Benedé, J. L., Chisvert, A., Salvador, A., Moreno-Garrido, I., & Blasco, J. · PLoS ONE 8(6):e65451 · 2013 · doi.org/10.1371/journal.pone.0065451Peer-reviewed
12. Scientific advice on titanium dioxide (TiO2) (SCCS/1661/23)Scientific Committee on Consumer Safety · European Commission, SCCS/1661/23, adopted 13 May 2024 · 2024 · health.ec.europa.euGovernment
13. Small amounts of zinc from zinc oxide particles in sunscreens applied outdoors are absorbed through human skinGulson, B., McCall, M., Korsch, M., Gomez, L., Casey, P., Oytam, Y., Taylor, A., Kinsley, L., & Greenoak, G. · 2010Peer-reviewed
14. Zinc absorption from a zinc oxide sunscreenGulson, B., Wong, H., Korsch, M., Gomez, L., Casey, P., McCall, M., McCulloch, M., Trotter, J., Stauber, J., & Greenoak, G. · 2012Peer-reviewed
15. Should nanoparticles be used in sunscreens?Gulson, B., McCall, M., Korsch, M., Gomez, L., Casey, P., Oytam, Y., Taylor, A., McCulloch, M., Trotter, J., & Kinsley, L. · Proceedings of the 18th International Conference on Heavy Metals in the Environment (ICHMET), Ghent · 2016 · doi.org/10.21825/ichmet.71265Conference
16. Support for the safe use of zinc oxide nanoparticle sunscreens: lack of skin penetration or cellular toxicity after repeated application in volunteersMohammed, Y. H., Holmes, A., Haridass, I. N., Sanchez, W. Y., Studier, H., Grice, J. E., Benson, H. A. E., & Roberts, M. S. · Journal of Investigative Dermatology 139(2):308-315 · 2019 · doi.org/10.1016/j.jid.2018.08.024Peer-reviewed
17. Stratum corneum is an effective barrier to TiO2 and ZnO nanoparticle percutaneous absorptionFilipe, P., Silva, J. N., Silva, R., Cirne de Castro, J. L., Marques Gomes, M., Alves, L. C., Santus, R., & Pinheiro, T. · 2009Peer-reviewed
18. Human epidermal zinc concentrations after topical application of ZnO nanoparticles in sunscreensKhabir, Z., Holmes, A. M., Lai, Y. J., Liang, L., Deva, A., Polikarpov, M. A., Roberts, M. S., & Zvyagin, A. V. · International Journal of Molecular Sciences 22(22):12372 · 2021 · doi.org/10.3390/ijms222212372Peer-reviewed
19. Zinc oxide particles and zinc ions in human epidermisHolmes, A. M., Song, Z., Moghimi, H. R., & Roberts, M. S. · ACS Just Accepted manuscript · 2016Peer-reviewed
20. Zinc from nano zinc oxide sunscreen in human skinHolmes, A. M., Kempson, I. M., Turnbull, T., Paterson, D., & Roberts, M. S. · 2020Peer-reviewed
21. Dermal absorption of nanomaterials titanium dioxide and zinc oxide based sunscreen: role of size and surface coatingBeer, C., Stenderup, K., Wang, J., Sutherland, D. S., & Nyengaard, J. R. · Environmental Project No. 1736, Danish Environmental Protection Agency · 2015 · www2.mst.dkGovernment
22. Titanium dioxide nanoparticle penetration into the skin and effects on HaCaT cellsCrosera, M., Prodi, A., Mauro, M., Pelin, M., Florio, C., Bellomo, F., Adami, G., Apostoli, P., De Palma, G., Bovenzi, M., Campanini, M., & Larese Filon, F. · International Journal of Environmental Research and Public Health 12(8):9282-9297 · 2015 · doi.org/10.3390/ijerph120809282Peer-reviewed
23. Safety evaluation of sunscreen formulations containing titanium dioxide and zinc oxide nanoparticles in UVB sunburned skinMonteiro-Riviere, N. A., Wiench, K., Landsiedel, R., Schulte, S., Inman, A. O., & Wang, J. · 2011Peer-reviewed
24. UVB exposure enhanced the dermal penetration of zinc oxide nanoparticles and induced inflammatory responses through oxidative stress mediated by MAPKs and NF-κB signaling in SKH-1 hairless mouse skinPal, A., Alam, S., Chauhan, L. K. S., Saxena, P. N., Kumar, M., Ansari, G. N., Singh, D., & Ansari, K. M. · Toxicology Research 5(4):1066-1077 · 2016 · doi.org/10.1039/c6tx00026fPeer-reviewed
25. Effect of flexing and massage on in vivo human skin penetration and toxicity of zinc oxide nanoparticlesLeite-Silva, V. R., Sanchez, W. Y., Studier, H., Liu, D. C., Mohammed, Y. H., Holmes, A. M., Ryan, E. M., Haridass, I. N., Chandrasekaran, N. C., Becker, W., Grice, J. E., Benson, H. A. E., & Roberts, M. S. · European Journal of Pharmaceutics and Biopharmaceutics · 2016 · doi.org/10.1016/j.ejpb.2016.04.022Peer-reviewed
26. Nanoparticles skin absorption: new aspects for a safety profile evaluationLarese Filon, F., Mauro, M., Adami, G., Bovenzi, M., & Crosera, M. · 2015Review
27. Review of environmental effects of oxybenzone and other sunscreen active ingredients, and inorganic UV filters zinc oxide and titanium dioxideSchneider, S. L., & Lim, H. W. · Photodermatology, Photoimmunology and Photomedicine 35:442-446 · 2018 · doi.org/10.1111/phpp.12439Review
28. Skin barrier function in infants: update and outlookRahma, A., & Lane, M. E. · Pharmaceutics 14:433 · 2022 · doi.org/10.3390/pharmaceutics14020433Peer-reviewed
29. Sunscreen drug products for over-the-counter human use (proposed rule)U.S. Food and Drug Administration · Federal Register 84(38):6204-6275, 26 February 2019, Docket No. FDA-1978-N-0018 · 2019 · www.govinfo.govRegulation
30. Primary sunscreen monographHealth Canada, Natural and Non-prescription Health Products Directorate · Natural Health Products Ingredients Database, monograph dated 27 February 2026 · 2026 · webprod.hc-sc.gc.caRegulation
31. Regulation (EC) No 1223/2009 on cosmetic products (recast)European Parliament and Council of the European Union · Official Journal of the European Union, L 342, 22.12.2009, p. 59-209 · 2009 · eur-lex.europa.euRegulation
32. Technically avoidable heavy metal contents in cosmetic productsFederal Office of Consumer Protection and Food Safety (BVL) · Journal of Consumer Protection and Food Safety 12:51-53 · 2017, online 6 October 2016 · doi.org/10.1007/s00003-016-1044-2Peer-reviewed

Update history

No substantive edit history is available in this build. The full commit record is available in git.