In most cases UVnat compared to the UVart gave qualitatively similar results. However, UVnat, with a lower fluence rate than UVart, gave similar yields of degradation. In addition, the fluence rate of the UVAart was higher than that of the UVAnat, which could be expected to degrade the sunscreens faster. But this is not the case, except for Sunscreen 4. Since the dose of UVAart was higher than UVAnat, Sunscreen 4 probably provides sufficient protection for the consumer.
Commercial sunscreens generally have low viscosity in order to be easy to apply. The temperature increase of the samples during UV exposure, especially after UVart, may lower the viscosity further. This may result in reductions of the optical path lengths of the samples. However, this was not the case in our study since samples kept on a heating plate for 20 min at 50°C showed a similar spectrum before and after heating.
Four of the seven sunscreens contain TiO2. If the particles are too small they may lose their scattering effect and, consequently, not give as good protection as larger particles. This may be the case for Sunscreen 5 (Fig. 3a). Several other studies show that inorganic chemical filters are not always photostable [6, 9–11]. Our study indicates the opposite, but seven sunscreens are a quite small amount of material, so this finding should be interpreted with caution.
When mixed with petrolatum, some sunscreens undergo degradation during exposure to UV radiation, especially in the UVA range . This is also the case for one of the most frequently used UV filters BMDBM. This compound is included in six of the seven sunscreens studied here (Table 1). Our results confirm the findings from other studies that sunscreens containing the combination of EHMC and BMDBM are photounstable, regardless of what other UV filters they contain [6, 9].
Some manufacturers of sunscreens claim that commercially available sunscreens are photostable because the photoactive species are in a vehicle that stabilizes them. This claim does not seem to be correct in several cases. There are several studies about how improvement of photostability may be obtained, e.g. with nanoparticle encapsulation of EHMC , liposphere preparation of BMDBM  or a combination with diethylhexyl syringylidene malonate and BMDBM . These findings are very interesting and will hopefully lead to an improvement in photostability in commercial available products.
When sunscreens without metallic oxide particles are compared, Sunscreen 1 seems to be more rapidly degraded than BMDBM dissolved in petrolatum. Not only does the UVA protection decline after exposure, but also the UVB protection. EHMC is one of the two UVB-absorbing filters present in Sunscreen 2, and the only one in Sunscreen 1. EHMC dissolved in petrolatum is rather photounstable . The vehicles of Sunscreens 1 and 2 are nearly identical. The UVA-absorbing compound benzophenone-3 (BZ-3) is added in Sunscreen 2. The presence of this compound may stabilize BMDBM, in agreement with earlier findings . Another stabilizer that may work is anisotrizine (CAS no 187393-00-6)is ; however, that compound was not included in any of the products in this study. Sunscreen 2 also has a higher SPF. However, a degradation manifesting itself in the UVA1 region should be noted.
Sunscreen 5 is photostable but does not contain any metallic oxide particles. This may be due to a vehicle that successfully prevents degradation and/or due to microstructures of the emulsion itself (Fig. 3a). It is interesting to compare this spectrum with that of Sunscreen 3 (Fig. 1c) which, according to the list of contents, includes TiO2 particles but does not show the scattering slope. The size of the particles may be too small (15 nm, according to the producer) to influence the absorption spectrum in the visible range. Small particles of TiO2 are expected to give maximal scattering in the UVB or UVC region. Larger particles can cause significant scattering also in the UVA and visible region. It follows that the small nanoparticles cannot give good protection in the UVA region in this case.
The peak between 350 and 375 nm in the absorption spectra of Sunscreens 3, 4 and 5 (Figs. 1c, 2, 3a) can be attributed to BMDBM. In view of this it should be noted that the UV exposure makes the products react quite differently. In Sunscreen 3 the BMDBM peak almost vanishes totally after 30 min of UVnat, while the peaks in the other two sunscreens are more stable. We suggested above why there could be degradation in the UVA range in Sunscreen 3 despite the presence of TiO2 particles. The reported stabilizing effect of 4-Methylbenzylidene camphor (MBC)  does not manifest itself in the case of Sunscreen 3.
The photostable Sunscreen 6 contains, in addition to BMDBM and TiO2, a third UVA absorber, terephthalylidene dicamphor sulfonic acid (TLDCSA), which can stabilize BMDBM. It has also been shown that TiO2 may stabilize ketoprofen and may be used in protecting photounstable species .
Many commercial sunscreens give, according to the manufacturers, good UVA and UVB protection. However, the photostability of the sunscreen in the UVA range is not always adequate. Most sunscreens offer good protection against UVB while the UVA photostability of some products decreases substantially during UV exposure. The potential toxicity of the photoproducts also needs to be investigated further.
For the consumer it is very difficult to know what product to choose, since the photostability varies between different brands and the photostability is not marked on the bottle. To know which photoactive compound the sunscreen contains is not good enough. The stability also depends on factors like preservatives, oxygen radical scavengers, and base formulation. It is not reasonable that the ordinary consumer should have knowledge of this. If the product claims to give broadband protection, this protection should remain also after sun exposure. The fact that sunscreens are photounstable has been known for many years. Our study clearly shows that there are still many photounstable products on the market. When buying a sunscreen, the consumer should automatically receive a photostable product.