Contributed by PETRA HUBER*, SARAH MEROLA, ZHAW Zurich University of Applied Sciences, Life Sciences and Facility Management, Campus Grüental


There is increasing interest in quantifying the penetration of active substances into the skin. Toxicologists wish to demonstrate the non-penetration of sunscreens to confirm product safety. Specialists in biopharmaceutical issues are interested in the penetration kinetics of active ingredients in various skin layers, the release of active ingredients from their carriers (“delivery”) and how different carriers such as simple solutions, emulsions or other carriers affect the efficacy of the product (1). In cosmetics both the penetration into the skin and efficiency of delivery of active ingredients are of interest. There is also a commercial need for better human skin equivalents (HSEs) as clinical skin substitutes and as models for permeation and toxicity screening. Testing methods must confirm that the barrier properties of the HSEs are comparable to skin (2).


It has been demonstrated that some active ingredients penetrate the barrier of the stratum corneum of the skin readily while others require the emulsion base in which they are incorporated to be modified. The penetration is determined by the molecular size of the substance and its lipophilic or hydrophilic character. The relationship between lipophilicity and hydrophilicity is represented by the dimensionless octanol-water partition coefficient, P. This describes the relationship of the concentrations of octanol (lipophilic) and water (hydrophilic) components detected in each phase of a two-phase mixture. The logarithm of the P value is positive for lipophilic and negative for hydrophilic substances. The penetration ability of a substance is described by its permeability coefficient (Kp).


Daniels (3) identified that a moderate lipophilic octanol-water partition coefficient (Pow between 10 and 100, corresponding to a logP between 1 and 2) and a molecular mass up to 500 g/mol (corresponding to daltons) indicate an ideal penetration of a substance through the stratum corneum of the skin epidermis.


As part of the drive to replace animal experiments, so-called Franz diffusion-type cells can be used in accordance with OECD Guideline No. 428 as an in vitro method to measure skin absorption (4). This offers an interesting alternative to the more costly in vivo method involving Raman spectrometric measurement on living skin (5).

Figure 1.png

The static test design uses a vertical Franz-type diffusion cell. As shown in Figure 1, the vertical glass Franz diffusion cell is composed of a lower compartment called a receptor and an upper compartment called a donor between which a membrane is positioned. The two compartments are held together with a clamp. The cells are surrounded by a water jacket to guarantee a constant temperature during the experiment. The OECD guideline recommends a temperature near normal skin temperature of 32 +/- 1 °C. The receptor chamber contains a receptor fluid that is constantly mixed by a stir bar. The receptor fluid should be chosen so that it is physiologically favorable for the skin and so that the test substance has an adequate solubility in it (4). If the solubility of the test substance in the receptor fluid is not sufficient, saturation and back diffusion may occur. Furthermore, the receptor fluid should neither influence the barrier function of the membrane nor the analytical procedure (6).


The donor chamber contains the test substance which is present as an aqueous solution or is formulated in such a way as to correspond to the way humans may be exposed to it. The test substance is applied on the skin (or membrane). The dose for percutaneous absorption should be infinite, where large volumes of active substances per unit area are applied (4).


The preparation of the individual Franz diffusion cells in the static model and the associated single-point measurement sampling, which in each case also means the termination of the experiment in the diffusion cell in question, are time and material intensive. The sampling of an adequate amount of receptor fluid is a delicate step. To maintain sink condition and to avoid air bubbles under the membrane, the fluid level should not change during the experiment. The sink conditions can be maintained by using multiple cells corresponding to the number of sampling points or by replacing the collected fluid with fresh medium. Both of these options are not optimal. By using a large number of cells the costs of material is extremely high. Furthermore, the usage of a large number of skin pieces (or skin equivalents) increases the heterogeneity of the skin samples resulting from the different skin sources and increases the standard deviation within multiple determinations. Alternatively, when replacing the collected receptor fluid with fresh medium unwanted errors may occur during the dilution step or when determining its concentration. In both cases a trained technician researcher needs to be present during the 24 hours of the experiment and a restricted number of sample points is possible.


An on-line measurement system that permits continuous measurement of penetration characteristics has been developed. The dynamical test design uses an in-line cell where the receptor fluid flows through the receptor chamber and simulates the blood circulation in the skin. The concentration of the test substance in the receptor fluid is determined directly by HPLC. This system capitalizes on the advantages of the Franz diffusion cell test method and can be utilized at the pre-screening stage to give time and cost-savings. This novel on-line measurement system is presented and discussed in this paper.






The Franz diffusion cell (Permegear Type 4G-01-00-25-20), with a custom built sample port of 3.5 mm, and the Franz diffusion cell V3 Stirrer were purchased from SES GmbH Analysesysteme (Germany). The synthetic membrane-based model, Strat-M®, for transdermal diffusion tests was purchased from Merck Millipore (Switzerland). Syringe filters (Chromafil Xtra PET 20/25) with a pore size of 0.20 μm were obtained from Machery-Nagel (Germany).


Caffeine (CA), (theobromine, TH, and epicatechin, EC, for the pre-tests) and phosphate buffer solution (PBS 1M and pH 7.4) were supplied by Sigma-Aldrich (Switzerland).


The UV/Vis spectrophotometer Cary 60, the fiber optic dip probe coupler and the fiber optic microprobe were purchased from Agilent (Switzerland). Cary WINUV software was used for every experiment. A second fiber optic microprobe (Falcata with a diameter of 3.2mm) was obtained from Hellma Analytics.  The main measurements, based on transflection, were conducted with this microprobe. Both fiber optic microprobes are developed to measure in small sample volumes (1 ml).


Diffusion method The diffusion study was performed using a Franz-type diffusion cell with a diameter of 25 mm, a surface area of 4.91cm2 and a custom made sample port of 3.5mm. The custom-made sample port was necessary to position the fiber optic probe in the acceptor chamber. The temperature of the cell was held constant by a thermostat-controlled water bath at 32+/-1°C. A small magnetic stir bar was introduced in the acceptor chamber to stir the acceptor fluid. The phosphate buffer solution was degassed for 30min with a water-jet vacuum pump and diluted with ultrapure water to 0.02 M. Subsequently the PBS solution was filtered with a syringe filter. Before filling the cell with the PBS solution, the fiber optic probe was positioned in the acceptor chamber so that no air bubbles were formed in the optical path during the filling. The sealing compound was placed on the surrounding of the acceptor and donor chamber. The Strat-M membrane was placed on the acceptor chamber avoiding air bubbles under the membrane and closed with the donor chamber and the clamp. After 15 minutes of equilibration 4ml of the donor solution with a concentration of 1g/L, which corresponds to an infinite dose of 4000μg, were applied and the measurement was started. During the kinetic study, measurements were taken every minute for the first hour and then every 30 minutes until the end of the experiment.



figure 2.pngThe penetration ability of a substance is described as its permeability coefficient (Kp [cm/s]). A comparison of the diffusion rates on a synthetic membrane (Strat-M®, Merck Millipore) showed that the selected pure compounds, Epicatechin (EC), Theobromin (TH) and caffeine (CA), applied at “infinite dose” (i.e. with a constant amount of active substance on the skin surface) had very different penetration abilities. With this pre-screening on-line method, a continuous permeation amount of caffeine from an aqueous solution (0.1%) was observed with a Kp of 9.07 *10-7 cm/s, see Figure 2. Because typically artificial membranes, compared with skin, do not represent such significant barriers, the lag times tL are very small and can be neglected (7). If barrier characteristics are given, for example with skin, the kinetic evaluation of the permeation coefficient should also include tL. This is the time required for diffusion through the skin. The investigation done on the synthetic membrane showed on a 95% significance level, the Pearson’s correlation coefficient exceeds r= 0.9986 that the growth curve is highly linear, i.e. the process is most probably of order zero. Howes et al. (6) proposed different mathematical models to evaluate/model/fit the resulting kinetic. This measurement system would enable a time-saving comparison of the barrier properties of various membranes; however, this was not the aim of the current study.


figure 3.pngMeasurements with both models of the microprobes demonstrated that the limit of detection (LoD) for caffeine was 3*10-6g/L and the limit of quantification (LoQ) was 1*10-5g/L. The calibration curve, measured at the wavelength λ 273nm  for caffeine, showed a linear relationship between the concentrations of 0.002 g/L (abscissa 0.1) und 0.018 g/L (abscissa 0.9)  validating the test method by microprobes within this range (see Figure 3). To check the accuracy of these data they were verified with the UV-absorbance detection by a micro plate assay (not shown).


In the method using UV-spectrometry (and a microprobe via transflectance), it is critical that the anticipated result falls within the linear measurement range, in accordance with the Beer-Lambert law, ideally with a transmission value of between 0.1 and 1. This expected value obtained from the acceptor medium in the Franz cell should be reviewed prior to the experiment with consideration given to the barrier properties of the selected membrane or in relationship either to the octanol-water partition coefficient, P, or the molecular mass, as stated above. In this study, neither the EC nor TH yielded results within the linear measurement range. This was probably due to their reduced solubility in the chosen medium and partly by their logP values rather than their molecular mass, see Tab.1. In general, the penetration of polyphenols is improved with smaller molecular size and moderate hydrophilicity (negative logP) according to Zillich et al. (7). If necessary, the initial concentration of the test substance in the donor medium should be adapted.






The in-vitro penetration studies with synthetic membranes in static Franz cells and an on-line detection by microprobe (optical principle of transflectance) were tested in triplicate for reproducibility. During one investigation the standard deviation within the results of +/-10% was due to the fragility of the setting of the Franz Cells and neither to the integrity of the membranes nor to the instability of the detection line. Measurement, for example by TEWL devices (transepidermal water loss), to proof the integrity of synthetic membranes is not specially required as it is with biological materials. However, this would make sense in the case of a resulting incoherent kinetic curve. For better readability, standard deviations have not been included in Figure 2 although the data was derived from multiple experiments.


To obtain consistent results, attention must be paid to the factors influencing degassing and filtration of the acceptor medium and to the absence of air bubbles under the membrane.


The all over recovery for the aqueous 0.1% caffeine solution was 99.3% +/- 4.4% (1.7% in the acceptor and 97.6% in the donor compartment) after triplicate analyses by UV-Vis spectrophotometry. The measuring principle was selective, i.e. there was no interference from the other substances in the media used in this study.  Individual substances could be detected at specific wavelengths. However, before tests are carried out on specific cosmetics, it must be established that auxiliaries are not detected within the selected wavelength range.


The above-mentioned properties and the determined limit of detection confirm the validation of this measuring method as described in the method section for single substances in aqueous solution.


In recent years, even though it is cost intensive, Raman spectroscopy has been established  as an alternative non-invasive method to directly determine penetrated active substances in human skin in vivo by focusing a laser into the top layers of skin (stratum corneum and within the first 200 μm of the skin) and recording the scattered Raman signals. Since this is not a standard method of measurement many degrees of freedom ensue. Hence, the suitability of Raman spectrometry for a specific measurement must be checked and validated. The limit of detection for measurement by HPLC can often be  lower compared to Raman spectroscopy provided that the substance to be detected is Raman active. Furthermore, absorption studies with quantitative recovery, which measure permeation and resorption processes in the in-vitro model, are not currently realizable.

figure 4.png

The on-line measurement technique permits the penetration kinetics of a substance to be determined in a full-thickness skin model (pig ear skin or human cadaver skin) instead of on synthetic membranes. In these cases, the full-thickness skin must be prepared separately to permit a quantitative recovery, as shown in Figure 4. To affirm the integrity of the dermatomed skin  the  mean of the TEWL flux was 3.83 g/hm2 +/- 1.30. The cut off-value of ≥ 20 g/hm2 was determined by internal investigation injuring the dermatomed skin intentionally with a scalpel (mean out of 20 single measurements).


After 8 hours of application, the substance epicatechin (EC) remained mostly in the donor compartment and a very small amount was detected in the skin layer viable epidermis, but this was at the limit of detection. The recovery for the aqueous 0.05% EC was 83% +/-0.72% after triplicate analyses.  Since the main barrier function of the skin is located in the stratum corneum, no or little permeation was expected. The amount of EC that was not detected on recovery and could have been permanently bound to skin proteins or possibly have been destroyed as a result of oxidation (1).


To optimize the delivery of the active ingredients into the skin, the polarity of the active ingredient must first be compared with the polarity of skin and the other emollient components (see the concept of Relative Polarity Index (RPI) by Wiechers J. (9)). This simple method of formulation should be checked and optimised before further skin application tests are carried out. Penetration enhancers, in addition to the chosen emollient and its polarity, affect the delivery and the penetration behaviour into the stratum corneum (9,10). Internal studies have demonstrated that the proposed in-line method can be adapted for investigations into the effects of using different formulation types or encapsulation systems with time-saving benefits.


Complex mixtures of substances in the acceptor medium should continue to be analysed by HPLC as long as their absorption maxima are not significantly different. The microprobe can be set for a lead compound. Some software programs also permit a wavelength scan during on-line measurement.


The aim of this study was to develop an alternative on-line measurement system fitted to the classic penetration study with Franz diffusion cell to directly quantify UV detectable substances in the receptor compartment. The classic in vitro method requires several samples to be taken from the receptor fluid over 24 hours and for corrections for the change in volume of the receptor fluid to be made or a number of cells equalling the number of the sample points to be used.


In summary, taking into account the previously discussed points, the following benefits can be derived from in-vitro penetration studies with a microprobe:


  • The combination of synthetic membranes permits prescreening of the fundamental penetration kinetics or release/delivery kinetics of substances.
  • Continuous monitoring of penetration kinetics can occur thanks to a theoretically unlimited number of measuring points. The duration of the experiments is based on the penetration behavior of the substance to be examined and the durability of the biological membranes (24-48 hours). The latter does not apply to synthetic membranes.
  • The presence of a technician researcher is not required and so the length of the experiment and the sample point can be set freely.
  • Due to the continuous measurement, less membrane material is required than with single-point measuring systems where the membrane is discarded after each data point.
  • The measurement system enables a time-saving comparison of the barrier properties of various membranes.




We thank Dr. Norbert Fischer, Vasilisa Pedan, Samantha Peters and Stella Cook, ZHAW, Wädenswil (CH), for their valuable support.





(1) Huber P., Merola S., Pedan V., et al., Study of polyphenol penetration from organic-aqueous cocoa extracts – antioxidant activity, IFSCC Conference 2015 Zurich, Proceedings, Full Paper and Poster, 2015

(2) Zheng Zhang, Bozena B. Michniak-Kohn, Tissue Engineered Human Skin Equivalent),  pharmaceutics 2012, 4, 26-41 Corp. 23-35, 2008

(3) Daniels R., Knie , U., Galenics of dermal products – vehicles, properties and drug release, JDDG 5:367–383, 2007

(4) OECD , OECD Guideline (428) for the testing of chemicals. Paris: OECD, 2004

(5) Huber, P., Adlhart, C., Luginbühl, V., Opitz, S., Morf, F., Yeretzian, C., Coffee based polyphenols with potential in skin care: Antioxidant activity and skin penetration assessed by in vivo Raman spectroscopy. Household and Personal Care Today, 9, 3. 60-65, 2014

(6) Howes, D., Guy, R., Hadgraft, et al. 'Methods for Assessing Percutaneous Absorption - The Report and Recommendations of ECVAM Workshop 13', Alternatives Lab. Anim., (24), 81-106,  1996

(7) Zillich, V. O., Schweiggert-Weisz U., HasenkopfK., et al.,  Release and in vitro skin permeation of polyphenols from cosmetic emulsions, International Journal of Cosmetic Science, 1–11, 2013

(8) ChemSpider ID: 393368s’ ACD/PhysChem Suite, Advanced Chemistry Development, Inc., Toronto, On, Canada,, 2014

(9) Wiechers J. W., Formulating for Efficacy Skin Barrier, Chemistry of Skin Delivery in: Wiechers J. W. (Ed.), Skin Barrier: Chemistry of Skin Delivery Systems, Carol Stream: Allured Publishing, 2008

(10) Barry B.W., Bennet S.L., effect of penetration enhancers on the permeation of mannitol, hydrocortisone and progesterone through human skin, J. Pharm. Pharmacol. 39 535-546, 1987


Original article from Teknoscienze, PETRA HUBBER, HPC Today, Vol. 10(5), pp. 40-43, 2015




“The Nexus Blog” is a sister publication of “The Nexus” newsletter. To sign up for the newsletter, please email, or if you have an ACS ID, login to your email preferences and select “The Nexus” to subscribe.


To read other posts, go to Green Chemistry: The Nexus Blog home.