Last Updated on August 16, 2026 by ashley
Can you make your own sunscreen? You can make a product that contains a UV filter. You cannot know what SPF it delivers. Those are two different statements, and everything difficult about this subject lives in the gap between them.
This is not gatekeeping. If you have built a stable emulsion, matched a preservative system to your pH, or made a shampoo bar that does not crack, you have real formulation skill. Sunscreen is not harder to make than those. It is harder to verify. And with a sun protection product, verification is not a finishing step. It is the product.
So this article will not hand you a zinc oxide recipe. It explains where the SPF number actually comes from, why it cannot be reasoned out from an ingredient list, and what happened when researchers took fifteen internet recipes into a laboratory and tested them properly.
Table of Contents
The number is not in the ingredients
Most formulation knowledge transfers cleanly between products. Add niacinamide at 5% and you have a 5% niacinamide serum. Use 3% of an emulsifier and you know roughly what it will do. Concentration goes in, performance comes out.
UV filters break that habit, and zinc oxide shows it best.
In 2008, a research group at the University of Nantes published in vitro work concluding that no commercial grade of zinc oxide they tested could reach an SPF above 10, even at 25%, the maximum permitted level. That finding spread widely. It still shapes a belief you will hear today: that zinc oxide alone cannot deliver high protection.
Today there are sunscreens on shelves carrying SPF 50 and SPF 50+ with 25% zinc oxide or less, established by the in vivo gold standard method.
Same filter. Same maximum concentration. A completely different answer.
Two things changed. First, zinc oxide grades improved, and so did the techniques for dispersing them. A 2024 review in Photochemical & Photobiological Sciences describes this directly: a decade of progress in dispersing zinc oxide properly has raised the SPF performance of zinc-based sunscreens, and newer particle grades now deliver more protection per percent used.
Second, part of that original low figure turned out to be a measurement artifact rather than a property of the material. We will come back to that, because it is one of the more useful things a formulator can understand here.
The conclusion, though, arrives early:
There is no percentage that reliably produces a given SPF. Not across filter grades, not across vehicles, and not even across decades for the same raw material.
Which contradicts something you have almost certainly seen. Percentage tables circulate constantly in DIY and small-brand communities. 5% zinc oxide gives you SPF 2 to 5. 20% gives you SPF 20 and above. They appear in workshop handouts, supplier literature and thousands of recipe posts, presented as settled chemistry.
They are not chemistry. They are a rough historical generalisation, drawn from particular materials in particular bases. Move any of those variables and the table stops describing reality. You can hit the stated percentage exactly and land well below the implied number, with no way of knowing.
A better mental model: this is not measuring salt into a recipe, where the amount decides the outcome. It is closer to asking how much light a coat of paint blocks. The pigment matters. So does how thickly it goes on, how evenly it spreads, whether it beads or films, and whether it stays put. You can know exactly what is in the tin and still not know what happens on the wall.
What actually determines SPF: the film
SPF measures how much ultraviolet radiation gets through the layer of product sitting on skin. That framing matters. Filters are not protecting skin in some abstract chemical sense. They intercept photons, and they can only intercept the photons that reach them.
So the geometry of the layer does at least as much work as its composition.
A 2014 study in the Journal of Biomedical Optics tested this. The researchers measured the real thickness of a sunscreen film after application by comparing surface topography before and after spreading, then ran the same UV filter system through five different vehicles: an oil-in-water cream, an oil-in-water spray, a gel, a water-in-oil emulsion, and an alcoholic spray.
The vehicle changed the result significantly. Low viscosity formulations left a thinner average film and measured a lower SPF. The formulation with a continuous oil phase left the thickest film and measured highest. Even how firmly the product was pressed on, and how long it was spread for, moved the number.
A follow-up from the same group calculated SPF from the measured thickness distribution and compared it against laboratory results. The two agreed closely. Their conclusion: the variation in SPF between formulations comes primarily from film forming properties.
Why thin spots decide the outcome
A sunscreen film is never uniform. Skin has texture, pores, fine lines and hair, and product settles unevenly across all of it.
UV attenuation through that film is exponential, not linear. Adding thickness to an already thick region buys almost nothing, because almost nothing was getting through. But a region that ends up half as thick lets through disproportionately more. The thin patches set the number.
The 2024 review puts it plainly: on skin, SPF is effectively determined where the film is thinnest and transmission is highest, while the product pooled in deeper creases transmits very little and contributes little.
This is why thickness distribution matters more than average thickness. Two products with the same filter loading and the same average coverage can still test differently. Radiation finds the gaps.
What that means at your bench
Every decision affecting how a product spreads, sits and stays is a decision affecting SPF.
Change the emulsifier and you change droplet size, film continuity and how filter particles distribute through the layer. Change the emollient blend and you change wetting and spread. Move from oil-in-water to water-in-oil and you change the whole film architecture. Thin the product for a nicer skin feel and you lay down less material.
None of this shows on a percentage table. All of it moves the number.
It is also why software that estimates SPF from filter type and concentration stays a screening tool. The 2014 researchers ran that calculation themselves and noted it could not predict the effect of formulation. It gets you a range, not a value you can print.
Dispersion, density and the water that leaves
Three physical effects deserve their own section, because none of them appear in a formula at all. They are where an experienced formulator is most likely to be caught out.
Agglomerates do less work
Inorganic filters are particles. They are suspended, not dissolved, and how well they are suspended is a processing outcome rather than a recipe line.
Zinc oxide particles want to stick to each other. When they clump, a cluster behaves optically like one larger particle instead of many small ones, and the surface area available to interact with UV drops. The label still says 20%. The film still contains 20% by weight. But much of that weight is no longer positioned to do the job.
This is exactly why the field moved forward. The 2024 review is explicit that improved dispersion technique, alongside newer particle grades, is what allowed formulators to get more SPF from each percent of zinc oxide. The material did not become more protective. Its distribution did.
Achieving that at scale means high-shear homogenisation, pre-dispersed filter pastes, dispersants matched to the particle coating, and process control that holds shear and temperature steady batch to batch. A stick blender does not do this. Neither does a domestic overhead stirrer. Particles look dispersed to the eye long before they are dispersed in the way that matters.
I spent a long stretch on this myself, working through mineral sunscreen trials with both a powder-grade zinc oxide and a pre-dispersed liquid grade of the same material. What stayed with me was not the difficulty of making a stable, pleasant product. That part went fine. It was that the measured protection kept moving as I changed the emulsifier system and the processing, while the filter percentage sat exactly where it was. I could shift the number without touching the ingredient supposedly producing it.
The density trap
Here is that measurement artifact, and it reframes what “more filter” means.
Sunscreens built on soluble organic filters have a residual oil phase with a density near 1.0 g/ml. Formulations loaded heavily with zinc oxide are much denser, roughly 1.3 to 1.7 g/ml.
Testing applies product by weight, in milligrams per square centimetre. A denser product occupies less volume at the same weight, so it spreads into a thinner film. And thinner film means lower measured protection.
A 2012 paper calculated that zinc oxide sunscreens at 20 to 22% would measure 41 to 44% higher if tested at the same film thickness as organic-filter products. The 2024 review goes further: a 25% zinc oxide emulsion can leave a film on the test plate as thin as 3.5 micrometres, which is less than the roughness of the plate itself. At that point a continuous film is not really forming.
That is worth reading twice, because it is genuinely counterintuitive. Pushing the zinc oxide percentage higher makes the product denser, which at a fixed application weight makes the film thinner. More filter does not linearly buy more protection, and past a point it can work against you.
Water leaves and the film thins
There is a second thinning effect, specific to aqueous systems. In an emulsion, roughly half the applied product evaporates after application. What remains is the film doing the protecting, and it is substantially thinner than what you spread on.
Anhydrous systems do not have this problem. There is nothing to evaporate, so the film you apply is close to the film that stays.
This is a large part of why balms and sticks tend to perform better per percent of filter than lotions built on the same loading. Not superior chemistry. A film that survives.
It also means two products with identical filter percentages, one anhydrous and one an emulsion, are not comparable in any useful sense.
The film is an emulsion problem before it is a sunscreen problem.
Emulsifier selection, droplet size, phase behaviour and processing are what decide how any emulsion sits on skin. My Emulsion Face Cream Formulation Guide covers that architecture in full: the emulsifier decision tree for O/W and W/O systems, complete formulas with exact percentages, and troubleshooting for when a batch behaves differently from the last one.
Even the laboratories are still arguing
One last piece, and it closes the technical case.
SPF measurement is hard under professional conditions. The in vivo standard needs a minimum of ten human subjects, rising to twenty-five, to land a result within plus or minus 17%. That is the tolerance inside one laboratory following one method.
Between laboratories it widens. When reference samples were tested across roughly a dozen laboratories using the same method, the standard deviation reached 30%.
And the debate about high-zinc products is live right now. That 2024 review exists because newer in vitro standards, developed largely on conventional emulsions with organic filters, systematically underestimate the in vivo SPF of high-zinc formulations, on average by about half. Dutch regulators re-tested commercial mineral sunscreens in 2024 and found the same pattern.
So the honest state of the field is this. After fifty years of standardisation, specialists with calibrated spectrophotometers and human volunteer panels are still refining how to measure the exact category of product that home formulators most want to make.
That is not a reason to doubt your own skills. It is a reason to be clear about what a number means. If professionals need ten to twenty-five volunteers and an accredited method to say “SPF 30” with a straight face, a batch made at home does not have an SPF. It has an unknown.
What happened when researchers tested the recipes
Everything above is mechanism. This is outcome, and it is short, because the findings are not ambiguous.
Fifteen recipes, made exactly as written
In 2021, a team published a study in the Journal of Cosmetic Dermatology titled “Demonstration of the dangerous nature of ‘homemade’ sunscreen recipes.”
Their method was fair. They searched for sunscreen recipes online, selected fifteen, and made each one in the laboratory following the published instructions exactly. Where a recipe gave quantities in tablespoons and teaspoons, they used tablespoons and teaspoons rather than converting to laboratory measures. The point was to reproduce what a person at home would actually get.
Then they measured SPF, UVA protection factor and critical wavelength.
Three of the fifteen contained no UV filter at all. Not an inadequate amount. None. Anyone following those recipes would be wearing a moisturiser they believed was protecting them.
The other twelve all measured below SPF 6, the minimum a product must reach in Europe before it can legally be called a sun protection product.
Fifteen recipes. Zero usable results.
What partial protection leaves behind
A 2025 study in Photochemical & Photobiological Sciences took a different approach. Instead of measuring SPF, the researchers looked at biological damage.
They took a homemade recipe published by a wellness blogger: almond oil, coconut oil, shea butter, beeswax, red raspberry seed oil, carrot seed oil and zinc oxide. If you have read DIY sunscreen content, that list will look familiar. They applied it to human skin explants alongside a commercial SPF 50+ product and a plain base lotion, irradiated all three, and examined the tissue.
The homemade cream did something. It reduced UV-induced DNA damage and it reduced epidermal thickening. That is a genuine effect and worth stating honestly.
But it failed to protect against sunburn cell formation. The commercial SPF 50+ product protected against every parameter measured.
This is the finding to sit with. The homemade product was not useless. It was partially effective, in a way the wearer cannot perceive, and ineffective on the endpoint most directly tied to acute UV injury. There is no feedback loop. Skin does not tell you which category your product falls into until the damage is done.
How widely this circulates
A 2020 study in Health Communication, published as “Pinterest Homemade Sunscreens: A Recipe for Sunburn,” analysed 189 pins found under homemade sunscreen and natural sunscreen.
Of those, 95.2% presented homemade sunscreen positively, and 68.3% recommended recipes offering insufficient UV protection. A third made a specific SPF claim, with values ranging from 2 to 50.
The reach is the part worth noting. Pins in the sample had been saved an average of 808 times. One had been saved more than 21,700 times.
This is not a fringe misunderstanding. It is a well-distributed body of confident, specific, wrong instruction, much of it carrying a number to lend it authority.
One risk running the other way
Worth flagging, because it inverts the usual assumption.
Several essential oils are phototoxic. Rather than offering protection, they increase the skin’s reactivity to UV and can cause burns and lasting pigmentation at exposures that would otherwise be harmless. Citrus oils are the best known, but not the only ones.
An essential oil added to a sun product for fragrance, or in the belief that it contributes protection, can make the finished product worse than applying nothing. I have covered which oils are involved and at what levels in a separate guide to phototoxic essential oils.
The short version
SPF is not an ingredient you add. It is a measured property of a finished formula, produced by the film that formula leaves on skin, and moved by the emulsifier you chose, how well you dispersed your filter, how dense the product is, how much water evaporates out of it, and how it spreads under a real person’s fingers.
That is why percentage tables do not work. It is why fifteen recipes made faithfully in a laboratory all failed the legal minimum. It is why accredited laboratories still disagree about measuring high-zinc formulations.
None of it reflects on your ability as a formulator. It reflects on the category.
The good news is that the sun care conversation is far bigger than sunscreen, and most of it is open to you.
After-sun repair is the most underserved category in indie skincare, and everything it treats is describable: inflammation, a compromised barrier, elevated water loss, oxidative load. Antioxidant day care has a genuine mechanism that sits entirely apart from filtering UV, and it is where botanical oils rich in tocopherols and carotenoids finally get to do honest work. Pigmentation correction may be the largest opportunity of all, because a great many people buying sunscreen are not afraid of burning. They are afraid of dark spots, and the corrective half of that concern is fully formulatable.
Anhydrous formats deserve a particular mention. The film-forming behaviour that makes water-free systems perform well as sunscreen bases is the same behaviour that makes them good occlusive repair products. You get the benefit of the format without inheriting the claim. My anhydrous formulation guide covers that base chemistry.
One caution as you build any of these. Implied claims are still claims. A product named Sun Balm, an image of a beach, zinc oxide printed on the front of the pack, or an Instagram reply telling a customer the zinc “gives you some protection” will all be read as sun protection claims, whatever the label says. If you are working out where those lines fall, I have written about it in selling homemade skincare legally.
Build those products. Wear a tested sunscreen. Tell your customers to do the same.
Working out what your product line should actually contain?
I work with indie brands on exactly these decisions: which products to develop and in what order, what each one can legally claim in your target markets, and where the technical risk sits before you commit to a batch. Written deliverables, fixed pricing, no retainer.
References
- Couteau, C., Alami, S., Guitton, M., Paparis, E., & Coiffard, L. J. (2008). Mineral filters in sunscreen products: comparison of the efficacy of zinc oxide and titanium dioxide by in vitro method. Die Pharmazie, 63(1), 58–60. PubMed
- Osterwalder, U., Hubaud, J.-C., Perroux-David, E., Moraine, T., & van den Bosch, J. (2024). Sun-protection factor of zinc-oxide sunscreens: SPF in vitro too low compared to SPF in vivo — a brief review. Photochemical & Photobiological Sciences, 23, 1999–2009. Open access
- Film thickness frequency distribution of different vehicles determines sunscreen efficacy (2014). Journal of Biomedical Optics, 19(11), 115005. SPIE
- Calculation of the sun protection factor of sunscreens with different vehicles using measured film thickness distribution (2016). PubMed
- McCormick, P. G., & Tsuzuki, Z. (2012). In vitro testing of zinc oxide sunscreens. International Journal of Cosmetic Science, 34, 291–297. PubMed
- Couteau, C. et al. (2021). Demonstration of the dangerous nature of ‘homemade’ sunscreen recipes. Journal of Cosmetic Dermatology, 20, 1788. PubMed
- Uncovering the efficacy of a natural homemade sunscreen in protection from ultraviolet radiation (2025). Photochemical & Photobiological Sciences. PubMed
- Merten, J. W., Roberts, K. J., King, J. L., & McKenzie, L. B. (2020). Pinterest homemade sunscreens: a recipe for sunburn. Health Communication, 35(9), 1123–1128. PubMed
- ISO 24444:2019. Cosmetics — Sun protection test methods — In vivo determination of the sun protection factor (SPF). ISO
