

Sunscreen makers offer mineral and non-mineral formulations, as well as products that combine mineral and non-mineral active ingredients. Mineral formulations incorporate zinc oxide or titanium dioxide in nano- and micro-sized particles that can be toxic if they penetrate the skin, but they offer strong and lasting UVA protection that is rare in non-mineral sunscreens. Most studies show that these ingredients do not penetrate skin to reach the bloodstream, but research continues. Sunscreens with these ingredients constituted 30 percent of the market in 2010.
The most common ingredients in non-mineral sunscreens are octisalate (found in 59 percent), oxybenzone (52 percent) and avobenzone (49 percent). Oxybenzone can trigger allergic reactions, is a potential hormone disrupter and penetrates the skin in relatively large amounts. Some experts caution that it should not be used on children. Some 55 percent of sunscreens rated by EWG are non-mineral, and 15 percent combine both mineral and non-mineral active ingredients.
EWG reviewed the scientific literature on hazards and efficacy (UVB and UVA protection) for all active ingredients approved in the U.S. Though no ingredient is without hazard or perfectly effective, on balance our ratings tend to favor mineral sunscreens because of their low capacity to penetrate the skin and the strong and lasting UVA protection they offer.
Mexoryl SX (ecamsule) is another relatively low-risk, effective active ingredient that provides UVA protection, but it is sold in very few formulations. Tinosorb S and M appear to be both safe and effective but are not yet available in the U.S. For consumers who wish to avoid mineral products, sunscreens containing avobenzone are the best options (at the maximum allowable level of 3 percent for the best UVA protection). The potential hormone disrupters oxybenzone or 4-MBC appear to pose greater potential risks than other common sunscreens; EWG recommends against buying sunscreens that contain this compound.
Oxybenzone is found in more than half of the 814 beach and sport sunscreens in EWG’s 2012 database. Experts raise concerns about its use in sunscreens for children because of its ability to penetrate the skin and its association with allergic reactions and potential hormone disruption:
Since 2007, EWG has highlighted zinc oxide and titanium dioxide as among the best of the available sunscreen ingredients for American consumers. In the 1980s, mineral sunscreens made with these minerals were thick and pasty white. Today, sunscreen makers use particles of between 20 and 200 nanometers in size to create lotions that are nearly clear.
Products with zinc and titanium, sometimes called “mineral filters,” tend to rate well in EWG’s guide because they block harmful UV radiation without penetrating healthy skin. Unlike other common sunscreen chemicals, zinc and titanium are not allergenic, do not break down in sunlight and do not disrupt the body’s natural hormones. EWG’s review of more than a dozen in vivo and in vitro studies found little to no skin penetration for nanoparticle zinc oxide and titanium dioxide on healthy human, pig and rodent skins (EWG 2009).
EWG’s assessment of the comparative benefits of zinc and titanium sunscreens that might contain nanoparticles is not meant as an endorsement of all nanoscale products nor of the manufacturing processes. Studies suggest that nanomaterials are toxic to fish and other aquatic life in the environment and can damage organs if they are absorbed through the skin, lungs or gut and enter the bloodstream.
EWG urges consumers to avoid mineral-based sunscreens sold as powders or sprays because they could inhale nanoparticles, with unknown consequences. EWG urges manufacturers of mineral-based powder and spray products not to use nano-scale particles. Recent evidence suggests that people have likely ingested significant quantities of nanoscale titanium dioxide used as a whitener in food and toothpaste for more than a decade (Weir 2012).
Sunscreens are used frequently over large portions of the body, which heightens the concerns over exposure.
A number of sunscreen chemicals are known to permeate the skin. Laboratory studies, tests on volunteers and human biomonitoring research have detected these chemicals in the general population. In the United States, the widely used chemical oxybenzone (or benzophenone-3) has been detected in 96 percent of the population (Calafat 2008, Wolff 2007). A recent European study detected four common sunscreen chemicals in mothers’ milk, indicating potential for exposure to the developing fetus and newborns (Schlumpf 2008). Indications of human exposure underscore the need for the FDA and other regulatory agencies to carefully review the toxicity of widely used sunscreen ingredients.
With this in mind, EWG evaluated US-approved sunscreen chemicals for both their ability to block UV radiation and for toxicity. The table below summarizes human exposure and toxicity information for 14 US-approved sunscreens and for three ingredients that are pending approval by the FDA. Ingredients are flagged for the following concerns:
No ingredient is without concerns; EWG’s rating system for sunscreens takes into account the range of concerns and differences in the weight of the evidence for each active ingredient.
| Sunscreen chemical | Percent of U.S. sunscreens that contain it |
Exposure (skin penetration and biomonitoring) | Toxicity concerns |
| Sunscreens with highest concern for human exposure and toxicity | |||
| 4-Methylbenzylidene camphor (4-MBC) | FDA approval pending | Limited skin penetration (1%) in vivo 1-3. Detected in European mothers’ milk at low parts per billion levels 4 | Strong evidence of hormone disruption 5-7; 8 9 10; thyroid effects 5; behavioral alterations in female rats 11 |
| Benzophenone-3 (oxybenzone) | 52% | 1-9% absorbed according to in vivo skin studies 1, 2, 12, 13; detected in volunteers’ urine 14, 15 and in European mothers’ milk 4. Present in 96% of Americans’ urine 16, 17; higher maternal exposures are associated with a decrease in birth weight for girls and an increase in boys 18 | Hormone disruption 19-22; reproductive effects and altered organ weights in chronic feeding studies 23. High rates of photo-allergy 24.Limited evidence of altered birth weights and increased odds of endometriosis in women. |
| 3-Benzylidene camphor | FDA approval pending | Hormone disruption 8; in vivo effects—behavior and estrous cycling 11 | |
| Octyl methoxycinnamate (OMC) | 29% | Limited skin penetration in vivo <1% 12, urine 1, 2, 14 and in European mothers’ milk at low parts per billion levels.4 | Multiple estrogenic effects 5, 6, 19; 21. Thyroid hormone reductions 25; and hormone-mediated immune effects.26 Moderate rates of skin allergy. {Rodriguez, 2006 #2683} |
| Padimate O | 0.1% | Limited skin penetration.27 Detected in European mothers’ milk at low parts per billion levels.4 | Estrogenic effects 19, 28 8. Damages DNA 29: causes allergic reactions in some people. |
| Sunscreens with moderate concern for human exposure and toxicity | |||
| Octocrylene | 44% | Limited skin penetration in vivo 27. Detected in European mothers’ milk.4 | Slight to moderate skin irritation.30 |
| Ensulizole | 0.6% | Skin penetration measured in vivo, documented concentrations in urine.31 | Occasional photoallergic reactions reported.32, 33 |
| Homosalate | 47% | Limited skin penetration in vivo <1% 12. Not detected in European mothers’ milk.4 | Limited evidence of hormone disruption.8, 19, 22, 28. Toxic metabolites 34 |
| Sulisobenzone (Benzophenone-4) | 0.q% | Skin penetration measured 35, estimated at 1%.36 | Limited evidence of hormone disruption.37 |
| Zinc Oxide | 36% | Very limited skin penetration.38 Estimated at 0.4% in volunteers for nano- and conventional particle sizes. Unknown whether it is in elemental (harmless) or insoluble particle form (toxicologically harmful)39 | No photoallergy or hormone disruption. Skin cell study found zinc nanoparticles provoked oxidative stress and DNA damage 40. Coatings may reduce skin reactivity. Zinc inhalation causes lung inflammation.41 |
| Titanium Dioxide | 23% | Very limited skin penetration 38: penetration of hairless mouse skin 42: no skin penetration in min-pigs 43. | No photoallergy or hormone disruption. Probable carcinogen when inhaled 44. Inhaled nanoparticles reach organs, cross placenta and enter brain.45-47 Skin damage in vitro 48. |
| Sunscreens with lowest concern for human exposure and toxicity | |||
| Avobenzone | 49% | Limited skin penetration in vivo 27; and in vitro (0.8%) 3, 49 | No evidence of photoallergy or hormone disruption. |
| Mexoryl SX | Limited approval (4 formulations); broader FDA approval pending | Limited skin penetration in vivo (0.16%).50 | No evidence of hormone disruption. Rarely reported skin allergy, more often in children. 51 |
| Octisalate | 59% | Limited skin penetration in vivo <1%.12 and in vitro (~0.5%).52 | Rarely, allergic contact dermatitis.53 |
| Tinosorb M | FDA approval pending | Low skin penetration measured in vitro.54 | No in vitro hormone effects. Did not stimulate uterotrophic activity in vivo.55 Allergic reactions uncommon.56 |
| Tinosorb S | FDA approval pending | No in vitro hormone effects; did not stimulate uterotrophic activity.55 | |
Table references:
1 – Janjua 2004. 2 – Janjua 2008. 3 – Klinubol 2008. 4 – Schlumpf 2008. 5 – Seidlova-Wuttke 2006. 6 – Seidlova-Wuttke 2006. 7 – Schlumpf 2004. 8 – Schreurs 2005. 9 – Durrer 2007. 10 – Maerkel 2007. 11 – Faass 2009. 12 – Sarveiya 2004. 13 – Gonzalez 2006. 14 – Hayden 1997. 15 – Felix 1998. 16 – Calafat 2008. 17 – Wolff 2007. 18 – Wolff 2008. 19 – Schlumpf 2001. 20 – Schlumpf 2004. 21 – Schreurs 2007. 22 – Ma 2003. 23 – NTP 1992. 24 – Rodriguez 2006. 25 – Klammer 2007. 26 – Rachon 2006. 27 – Hayden 2005. 28 – Gomez 2005. 29 – McHugh 1997. 30 – Odio 1994. 31 – Vidal 2003. 32 – Berne 1998. 33 – Schauder 1997. 34 – SCCP 2007. 35 – Kurul 2001. 36 – Benech-Kieffer 2000. 37 – Molina-Molina 2008. 38 – Nohynek 2007. 39 – Gulson 2010. 40 – Sharma 2009. 41 – Sayes 2007. 42 – Wu 2009. 43 – Sadrieh 2010. 44 – IARC 2006. 45 – Takeda 2009. 46 – Shimizu 2009. 47 – Park 2009. 48 – Pan 2009. 49 – Montenegro 2008. 50 – Benech-Kieffer 2003. 51 – FDA 2006. 52 – Walters 1997. 53 – Shaw 2006. 54 – Mavon 2007. 55 – Ashby 2001. 56 – Gonzalez-Perez 2007. [click here for full reference list]