Skin penetration is measured in vitro by mounting a disc of human skin between two chambers of a diffusion cell, applying a finite dose to its surface, and quantifying what crosses into the fluid below over 24 hours. Regulators accept the result only under fixed conditions: 32 ± 1 °C, split-thickness membranes of 200–500 µm, and mass recovery of 85–115 %.
Published 22 September 2026. Last reviewed 22 September 2026.
What a diffusion cell is
A Franz cell is a two-chamber glass apparatus in which a membrane of excised skin separates an upper donor compartment, open to the air, from a lower receptor compartment filled with a stirred, thermostatted fluid. The test substance is applied to the skin surface; the receptor fluid is sampled at intervals; the amount recovered in that fluid is what the method calls absorbed.
The OECD codified the design as Test Guideline 428: Skin Absorption — In Vitro Method, published 22 November 2004 (doi:10.1787/9789264071087-en). The guideline accepts both static and flow-through cells, specifies application rates of roughly 1–5 mg/cm² for solids or up to 10 µL/cm² for liquids, and sets the standard observation period at 24 hours.
Two terms are worth defining before the numbers, because they are routinely conflated in marketing copy:
- Percutaneous penetration is the movement of a substance into the layers of the skin — stratum corneum, viable epidermis, dermis.
- Dermal absorption, in the regulatory sense, is the fraction that crosses the skin into the receptor fluid, taken as a proxy for entry into the systemic circulation.
They are not the same quantity, and the difference decides how a result should be read.
The conditions that make a result admissible
For cosmetic ingredients in the European Union, the operative document is the Scientific Committee on Consumer Products' Basic criteria for the in vitro assessment of dermal absorption of cosmetic ingredients (SCCP/0970/06), adopted at the committee's seventh plenary on 28 March 2006. Its requirements are specific enough to be checked.
| Parameter | Requirement (SCCP/0970/06) | Why it is fixed |
|---|---|---|
| Membrane | Human skin preferred; split-thickness 200–500 µm as the general rule; pig skin acceptable | Thickness governs the diffusion path length |
| Temperature | 32 ± 1 °C at the skin disc | Corresponds to normal human skin surface temperature; diffusion is temperature-dependent |
| Barrier integrity | Checked before use — tritiated water, caffeine or sucrose permeation, or TEWL / transcutaneous electrical resistance | A damaged membrane overstates absorption |
| Mass balance | Overall recovery of test substance and metabolites within 85–115 % | Unaccounted material makes the absorbed fraction unverifiable |
| Replication | Minimum six evaluable samples from at least three donors per dose; coefficient of variation below 30 % | Inter-donor variability in human skin is large |
| Duration | 24 hours, at least six post-application sampling points including one at 30 minutes | Captures both early flux and steady state |
| Stratum corneum | Material retained in the stratum corneum at sampling is not counted as dermally absorbed | It is assumed not to reach the systemic compartment |
Rodent skin is excluded from this framework for a reason the committee states plainly: it is two to ten times more permeable than human skin. A penetration figure generated in rat skin is not transferable to a human forearm.
What a real measurement looks like
Iliopoulos, Chapman and Lane applied 2 % (w/w) 3-O-ethyl-L-ascorbic acid in a propylene glycol / propylene glycol monolaurate / isopropyl myristate vehicle at a finite dose of 5 µL/cm² to heat-separated human epidermis in Franz cells. At 24 hours the cumulative amount permeated was 49.4 ± 4.1 µg/cm², corresponding to 58.0 ± 4.2 % of the applied dose. Permeation was first detectable at 5 hours in human skin. A reconstructed human skin equivalent run alongside gave 41.3 ± 2.0 µg/cm² (55.1 ± 1.8 %), with permeation detectable from 2 hours — statistically indistinguishable from human skin at 8, 10, 12 and 24 hours (p > 0.05), but not at the early time points (International Journal of Cosmetic Science 2021;43:107–112, doi:10.1111/ics.12675).
Three things in that paragraph deserve attention. The figure is a percentage of the applied dose, not of the amount in the bottle. It refers to one specific derivative in one specific vehicle — the solvent system is part of the result, not a neutral carrier. And the early-time divergence between the two membranes shows that two models can agree at 24 hours while disagreeing about the first five.
What the measurement does not tell you
It does not measure what stays in the skin. The regulatory calculation deliberately excludes stratum corneum content, because the question being asked is one of systemic safety. For a cosmetic active whose site of action is the upper epidermis, that excluded fraction may be the fraction that matters. The same experiment, read for a different purpose, yields a different relevant number.
It does not account for where on the body the product is used. Follicular density varies by an order of magnitude across human skin: 292–448 follicles/cm² on the forehead against 29–93 follicles/cm² on the back (Todo H., Pharmaceutics 2017;9(3):33, doi:10.3390/pharmaceutics9030033). The same review reports a correlation of r = 0.79 (p < 0.05) between human and pig skin permeability across eleven compounds — a reasonable surrogate, not an identity.
It does not measure stability. A molecule can traverse a membrane efficiently and still have oxidised in the jar beforehand. Penetration and stability are independent variables, and a formulation has to satisfy both. Our own open-access technical report on this point is deposited at doi:10.5281/zenodo.22491021.
How to read a penetration claim
A claim that an ingredient “penetrates the skin” carries almost no information until four things are stated: the membrane, the applied dose, the vehicle, and the time point. Where those are absent, the number is not checkable. Where they are present, it can be compared against the acceptance criteria above — which is the whole point of having them.
Related reading on this site: what determines whether a molecule penetrates skin, and why topical vitamin C is formulated below pH 3.5.
Scope
This article describes published findings and regulatory test criteria. It is a description of measurement methodology and does not constitute a product claim, nor a statement about the performance of any specific product. TTS Clinical formulates according to compounding-pharmacy criteria; nothing here should be read as a claim of therapeutic effect.