Congratulations, the newly developed formulation is near perfect, looks good and smells great. The sensorial panel has given it rave reviews. The “chemical-free” cosmetics scaremongers on the Internet will find it contains only wholesome ingredients. Manufacturing loves it because it almost makes itself, and the accountants cannot believe its low cost. Even better, it contains the company’s latest XY95 ingredient, proven in lab tests to do wonders to the user’s skin. The “9 out of 10 women” test has also been passed, where the old trick of a 10-day scientific “washout” period has been enacted, so that even the simplest formula would have had the panelists saying their skin feels much softer than before.
There is only one piece of information lacking before the company makes millions and a promotion is assured: the result. Unfortunately, an expensive clinical trial with a genuine comparison of the same formulation with and without the active shows no difference. All the improvements came from a placebo effect and those other two well-proven ingredients: oil and water. However the statisticians look at the data, it seems as though the wonder ingredient XY95 is having no effect whatsoever. What could have possibly gone wrong? All the tests were double-checked, so there can only be one conclusion: Hardly a molecule of XY95 made it into the right part of the skin. This scenario is played out countless times throughout the cosmetics industry—a fact known to industry insiders but for obvious reasons, is not loudly proclaimed—and is largely unnecessary. Much wasted effort could be avoided by thinking through the delivery issues beforehand.
Molecular weight (MWt): The first issue to consider is the MWt of the active. A rule of thumb adopted by the pharmaceutical industry, which has been confirmed many times in cosmetics is: Anything with a MWt greater than 500 stands no chance of being practically delivered via a conventional route. Indeed, 500 is part of Lipinski’s “Rule of 5,” at which point pharmaceutical researchers begin giving up hope for any sort of delivery. Some may disagree on how molecules diffuse through skin but there is no question that large molecules diffuse more slowly than small molecules.
The effect of MWt on penetration is probably in the region of squared or cubed; i.e., if a molecule of a MWt = 100 enters the skin in 30 min, then something with a MWt of 500 penetrates at a factor of 25 or 125 more slowly, somewhere between 12 hr and 2.5 days. A recent, more pessimistic view of a factor of 2,000 comes from the correlation of Magnusson et al.1 One’s biology may be different; there may be a specific enzymic system such as those involved with lipid-handling within the skin that pulls large molecules through; the skin enzymes may actively cut the molecule into smaller pieces that can then penetrate; or maybe this large molecule or nanoparticle is intended to penetrate via the follicles. All of these routes are fit for their purpose so long as the intended route is known in advance so the formulation can be optimized for it. Anything with a MWt greater than 500 stands no chance of being delivered via a conventional route.
Without such special delivery routes, a general formulation with a high MWt active will provide no benefit other than enabling the claim of: “Contains XY95, which is proven in lab tests to do great things.” However, regulatory bodies around the world realize that consumers read this as “Contains XY95, which in this formulation will do great things,” and are becoming less and less happy with such marketing tricks.
Solubility: The next issue to consider relates to general solubility. Is XY95 basically soluble in water or oil? If it’s water-soluble, well, good luck. Skin obviously has been designed to be a reasonable barrier to aqueous systems—to keep the good water-soluble things in the skin. So, in general, actives that are soluble only in water do little good, with which both the lipid-only brick-and-mortar model of skin, i.e., high log P is good, and the “skin is just a polymer with solubility characteristics” model of skin agree.2 There are, of course, great exceptions. Some water-soluble actives speed through the skin; caffeine is a good example, which does so for unknown reasons. Adding to the problems of rationally delivering aqueous actives beyond the stratum corneum (SC) is the fact that there is often little water remaining in a formulation shortly after it has been applied to the skin, as is discussed further.2
If the active is oil-soluble, then an approach such as the previously described Formulating For Efficacy (FFE) produces a good balance of cost, speed, feel and other trade-offs in a formulation, allowing for a richer use of various oil chemical structures rather than just relying on something monodimensional, such as log P.2, 3 The FFE approach is based on solubility. The first requirement is sufficient solubility of the active in oil, so that potentially irritating crystals of the active do not form on the user’s skin. The second requirement is that the oils be nicely soluble within the SC, which provides several benefits. For example, if the oils penetrate, then the active dissolved in them will, too. The more the oils penetrate, the more the skin swells with the oil, which naturally increases the diffusion rate because the diffusion is now partly through the oil; this in turn allows even more oil and active to enter. In this example, think of skin as a polymer. Without this virtuous cycle, diffusion through the skin is slow.
FFE characterizes solubility in terms of three Hansen Solubility Parameters,4 so the actives, oil and skin each have numbers from which their relative solubilities can be calculated. This means that when marketing requires formulators to change an ingredient in the oil to one with a different solubility, the formulation can be rationally adjusted to compensate. The FFE approach encourages a scientific way to think about the oil, which, for example, makes it obvious that an oil such as dimethyl isosorbide (DMI) has desirable features, i.e., is a good Hansen Solubility match to skin and many APIs, in terms of delivering many typical actives. Smart ingredient suppliers use such FFE principles to create molecules that exist in the right parts of the solubility space, to allow for more formulation options. However, the most effective scientific formula in the world, in terms of delivery, is not going to sell if people do not like the way it feels on their skin, so other ingredients must be added that offer no benefits for delivery but benefits for look and feel. Unfortunately, such ingredients tend to make delivery much more difficult.
For example, most skin formulations contain long alcohols attached to a long acid, such as cetyl oleate, lauryl palmitate, etc. These materials have two big disadvantages in terms of delivery. First, regarding solubility, they do not dissolve many of the desirable actives; i.e., their Hansen Solubility Parameters are very far from those of typical active pharmaceutical ingredients (APIs), and for the same reasons, they are not soluble in skin, thus they do not helpfully swell the skin. Second, these structures are so large that even if they were soluble in skin, they would diffuse too slowly to be much help; the squared or cubed dependence of diffusion coefficient is a powerful force against their migration. So while these alcohols attached to a long acid may be great for skin feel and act as a moisture barrier to help preserve skin hydration, they are positively bad for delivering actives. One reason that isopropyl myristate is commonly used is because it is on the borderline; the myristate structure is long, but the isopropyl part is small enough to allow some solubility and some mobility.
One objection to the FFE approach is that, as is generally known, most formulations, for reasons of cost, convenience and general expectations, are not delivered from oil formulations but from aqueous emulsions. However, the standard defense is: The first thing that happens when an emulsion is applied to skin is the water disappears after the first 15 min or so.4 So the actual formulation on the skin is the oil that has been so carefully optimized. There is a flaw in this argument discussed (see Emulsifiers), but it is true that on the whole, delivery takes place from the materials or stuff remaining on the skin once the first 15 min of changes has passed. Therefore, this stuff can greatly impact the delivery of actives, as will be discussed.
Referring to the materials that remain on skin after product application as stuff requires further explanation. This author has found a striking phenomenon in countless discussions about cosmetic formulations: many individuals do not think beyond what all the materials in a formulation are doing, especially in terms of delivering actives, once they are applied to the skin. Formulators may understand the role of each ingredient but the preservatives, for instance, are only thought of as preserving the formulation until the moment of application. The emulsifiers create a nice, stable emulsion and the emollients are largely there for the critical moments of “feel” as the formulation is applied to skin, with some long-term feel properties as well. Silicones are also usually there for the same moments of “feel,” and the water just delivers the rest of the formulation and, perhaps, adds hydration. Thinking about these individual components in the context of delivering actives can be enlightening—and discussed next, starting with water.
Water: Gone in 15 min?
What happens to the water in a formula when the product is applied to skin? This question is asked surprisingly rarely, and replies vary wildly. The general consensus seems to be that it has mostly evaporated after 10–15 min, when applied as a typical formulation in typical quantities on typical skin. The answer for any one specific case is probably the most important fact that any formulator can obtain with little effort: Take a glass slide, rub the formulation on it in a manner similar to applying it to skin, then place it on an electronic balance connected to a computer and monitor the weight lost with time. A temperature close to that of skin and an air flow relevant to the intended use are helpful refinements. Usually the loss of weight is linear with time, and most of the water is gone after 15 min.
If there is a sudden break in the weight-loss curve, it may well indicate a phase change in the system; for example, the emulsifiers change to a more closely packed phase that blocks water loss. This is a useful fact to know and is extra data gleaned from a simple experiment. Of course, water loss from skin will be faster since the skin itself will absorb some water but monitoring water on the skin is not as easy, so the data from the glass slide can be used as an upper limit for the time during which there is sufficient water to do anything useful if it is intended to keep the active in solution or to, somehow, aid delivery.
Assuming that most of the water has gone after 15 min, the formulation team now knows that if the active is mostly water-soluble, it will be stranded high and dry on the skin, probably going nowhere other than onto the user’s hair or clothes. Thus, if a good amount of water is required to keep the active soluble and deliverable, then the formulation should include plenty of extra stuff to keep the water from evaporating—not an easy task. If the active is intended for oil delivery, then the disappearance of water is a good thing. Indeed, any materials that stop the water from disappearing will probably slow the delivery of the active. This is important to avoid since in the meantime, the active can be lost via various other routes. This raises questions about some often overlooked delivery considerations.
How much is left to penetrate? If a test formulation designed neither to permeate nor evaporate is applied to someone’s skin, how much of that formulation will still be on the skin after 1 hr, 2 hr, 3 hr, etc.? This is an important question that has not received much attention despite the fact that it is simple to study using unsophisticated techniques.
Most of any such formulation will have been lost after a few hours simply by chance transfer to clothes and other nearby surfaces, so while vast efforts are spent in sophisticated confocal Raman work on actives penetrating under very controlled circumstances, little thought has been devoted to knowing how much of the formulation might still be around to penetrate the skin. In addition, since little thought has been given on quantifying the loss, even less has been spent on methods to avoid it. If most of a formulation is lost after the first few hours, then any experiments with Franz cells after 24 hr are totally irrelevant. Delivering an active in just a few hours is much more challenging than delivering it over a 24-hr period, but if a formulation is genuinely long-lasting, the chances of achieving efficacy are greatly enhanced.
How can one measure the remaining ‘stuff’? The following experiments can easily be performed in any formulation lab. Take two simple oil formulas without water or alcohol, to avoid evaporation complications, each including ingredients known to be essentially non-volatile. One formulation should be of a generally low MWt and the other, a generally high MWt. Find an area on a volunteer’s skin and spread equal amounts of each formulation to create equal areas, typically 2 mg/cm², being sure that nothing touches either surface for a few hours. Using FFE, one can do this experiment virtually; the diffusion process of the different components can be modeled based on their MWts and will show that after approximately 1 hr, most of the low MWt stuff has gone into the skin while most of the high MWt stuff remains on the surface.
The visual experiment confirms this—be sure to capture the data with a camera; a smartphone will work well enough. It will be difficult to see any trace of the lower MWt formulation even if it contained some high MWt components after an hour or so, and the higher MWt formulation, even if it contained some low MWt components, will look virtually unchanged. Imagine this same experiment but with the active in the oil; this FFE diffusion model demonstrates what is intuitively obvious: More of the active will enter the skin with the low MWt formulation, assuming that the solubility of the active in each of the formulations is sufficient. However, there is an important caveat about low MWt oils. Although a molecule such as DMI can be helpful for skin penetration, with a high MWt active, the DMI might be too small—i.e., it can rush through the skin while the large active struggles to keep up with it. The result is the active being stranded on the skin surface. By conducting these simple visual experiments, anyone can get a good idea of the right balance between the solubility time scales determined to be relevant to delivering the active.
The same experiment can also be performed for water-soluble actives. Again, forget the water and simply compare formulations that keep the active stable and soluble. A typical comparison might be between a mostly propylene glycol formulation, and one with large amounts of high MWt PEG. The PG formulation will disappear quickly, not because the skin especially likes PG, but because it is a small molecule that can diffuse quickly. The PEG formulation will remain unchanged on the surface. Whether either of these scenarios will deliver the active is an interesting question, but at least this simple experiment sets the scene for a good debate.
The opposite of water in formulations is silicone. In general, silicones are not nicely compatible with many of the other ingredients, especially the water, but their moment of glory is the feel they can impart when the formulation is applied to the skin. What was formerly a nuisance in formulations, i.e., their tendency to rise to the surface due to low interfacial tension, has become their strength during application. But what happens once they are applied? High MWt silicones simply remain on the skin, which can be bad or good. Bad because they add little to solubility—nothing much is soluble in most silicones, so they can block the active from getting to the skin. Good because if they migrate to the surface of the formulation, then they can stop the evaporation of things that are desired to remain, such as water, and might reduce formulation loss via random transfer.
Reduced transfer comes about because a silicone surface has low surface energy, so less of the formulation will stick to, say, clothes upon any chance contact. This idea is often used for overnight cosmetics, where an extra silicone layer is applied to the underlying moisturizer to keep it from adhering to bed clothes. This raises a question of how much overnight cosmetics have, in the past, been lost within the first few minutes of going to bed. Some silicones are also somewhat volatile, so if they are added just for the moment of application and required to be lost soon after to avoid what in some cases can be a “chemical silicone feel,” it would make sense to know how quickly they are lost. Again, however, it seems that the rate of loss of silicone is not high on anyone’s measurement agenda.
Surface Active Agents
A question often asked by manufacturers in relation to the FFE approach is, “What about the surfactants in the formulation?” The first answer to this, i.e., “They are only present at low levels of a few percent,” is true but as will be discussed later, this can be misleading. It is worth focusing here on another issue: The curious notion that surfactants, here meaning “surface active agent” and not referring to emulsifiers, intrinsically behave as surfactants simply because that is what they are called. Yet in the absence of water, they have no obvious way to behave as surfactants and so become simply high MWt molecules. So again, it is important to know how quickly the water disappears and whether the surfactants undergo significant phase changes during drying, which may or may not be detrimental to delivering the active. This transition from surfactant to general high MWt molecule also seems to be discussed very little. Anyone with a polarizing microscope and a bit of patience can quickly check the fate of an emulsion when it is applied to a glass slide. Of course, a cover slip must be placed on the sample for proper viewing, so the experiment requires multiple samples left to dry for different amounts of time. It is a simple experiment that often yields nothing interesting—i.e., the emulsion just dries out with no obvious change other than the disappearance of water. However, when a spectacular change appears in the polarization, announcing a phase change, the formulator gains important information that often is highly relevant to skin feel.
There is also a hazy notion, backed by constant assertion and a few relevant experimental facts, that the right surfactant would be great for skin delivery. One source for this idea comes from the trend for using liposomes and other –omes, all of which are based on surface active agents, to “aid in delivery.” The idea that the liposomes themselves somehow pass through the skin carrying their contents is nonsense; it does not survive two seconds of rational thought about the skin and possible delivery routes. There simply is not a route through which these relatively rigid particles can weave their way in and if they must break apart in order to work, they are no longer acting as -omes. As one review of a new type of –ome states, “Recently, it became evident that in most cases, conventional ‘rigid’ liposomes are of little or no value as carriers for transdermal drug delivery, as they do not deeply penetrate skin but rather remain confined to the upper layers of the SC.”5 They might be useful for holding an unstable active in the formulation up to the point of delivery, and this is an excellent reason for using them, but once the water is gone, they are just more stuff sitting on the skin.
Another reason that surfactants might be considered helpful for delivery comes from review papers that “prove” their efficacy; yet far too often, revisiting the original papers will show little support for such assertions. First, the effects in the papers are usually measured at infinite doses of the surfactant solutions in Franz cells, which means they have water present and are therefore, unlike the real situation, acting as surfactants on fully hydrated skin. Second, the effects are often trivial, measuring a difference of just a factor of two. Third, in surfactant terms, the effects are implausible; a C16EO2 might have the same effect as a C16EO10 despite the fact that as surfactants, these are wildly different molecules.
One of the few surfactants with proven ability to help things through the skin is sodium lauryl sulfate (SLS), and it does so in a most irritating manner. In general, surfactant molecules are much too large to do anything useful in relation to delivery. Of course, if they are similar to skin lipids, then the skin’s enzyme system can actively move them around, but it is still not clear why this should have any significant effect on delivery of the active.
For those who believe that surfactant molecules can zip through the skin carrying the active with them, just do a simple experiment. Put 2 mg/cm² of the relevant surfactant onto the skin and via simple surface-stripping tests measure how much is on the skin at 1 hr, 2 hr, 3 hr and later. If the test is done in a manner that avoids accidental transfer to clothes, it is a safe bet that more than 90% will still be there after a few hours.
The flaw mentioned earlier in the simple FFE argument can now be discussed. Recall that the argument said that the emulsifiers are present at only small amounts in any formulation, so in terms of the overall formulation, these molecules could be ignored once the water has gone. The mistake made was to not think in terms of overall stuff, which indeed was the mistake that originally spurred the writing of this article. If the oil phase containing the active is 10% of the overall formulation and the emulsifier is 5%, i.e., a small percent, then when the water has gone, the emulsifier is 33% of the formulation sitting on the skin. Assuming that the previous paragraphs are right and these molecules mostly sit on the surface, what effect will they have on the active? If the active is insoluble in them, it might crash out of the formulation—and a crystalline active is not going anywhere fast. Conversely, if the active is nicely soluble in them, then the active has a choice: stay where it is or diffuse into the skin. On the whole, the choice will be to stay where it is, i.e., the partition coefficient into the skin will be low, so only a small percentage of emulsifier in the total formulation can have a devastating effect on skin delivery when it becomes a large percentage of the formulation on the skin.
Similar issues that apply to the other stuff apply to preservatives. In a “paraben-free” formulation, for instance, there might be a relatively high percentage of a less effective preservative. Say that it was 1% of the overall formulation, which seems irrelevant; but once the water has gone, if the oil phase is 10% then that 1% preservative will be 10% of the formulation. The consequences of this would need to be thought through carefully, depending on whether the preservative is a large or small molecule. In fact, it is conceivable that in some formulations, the molecule most helpful for delivery of the active will be the preservative, as many of them are relatively small molecules and, like the parabens themselves, are relatively compatible with the skin in Hansen Solubility terms.
Finally, it follows from all of the above that any attempt to compare delivery systems in general will be doomed. To say that hydroethanolic systems are better or worse than microemulsions, w/o emulsions, o/w emulsions or -omes in general is ignoring the fact that for the majority of real-world systems, which are non-occlusive, after the first 15 min, there is no meaningful amount of the delivery vehicle remaining; just a lot of stuff, most of which will not be doing anything helpful and may even be getting in the way of the components that might help. Yes, the delivery vehicle is tremendously important to other aspects of the cosmetic such as feel, but once you become alert to this irrelevance of the general nature of the delivery vehicle, vast tracts of the skin delivery literature start to look as though the researchers may be asking entirely wrong questions.
So far, the tone of this article has been negative, but it is easy to switch to a more positive note. The great news is: Thinking about the issues of stuff is easy, all it takes is the formulation team getting into the habit of thinking through each component not only during its vital few seconds of fame, e.g., controlling feel during application, but also as that component sits on the skin while the active is being delivered. Further, the tests to determine the fates of components—i.e., mostly weighing experiments, sometimes with surface stripping—are very simple to perform and hugely informative.
There is one final massive bonus. Countless hours are spent on expensive skin permeation tests for actives using Franz cells. These not only soak up time and resources, they also are often grossly misleading; the skin samples used are only an approximation to the reality of a subject’s skin after application of the formulation. Also, while those who have confocal Raman spectrometers conduct difficult experiments mapping the concentration of the active in the SC of a human volunteer, there is a far simpler, more direct method that is rarely used. Place a small amount of the formulation onto a volunteer’s skin and at regular intervals take samples, e.g., by surface stripping. Place these samples into an HPLC and see how much active remains in the formulation above the skin. The results are often shocking, as the amount of active on the skin—when corrected for other losses such as transfer to clothes—will be close to 100% of the original, maybe locked inside some phase-changed emulsifier, of which everyone was unaware. In other words, do not worry about how much has penetrated the SC; instead, worry about the much easier problem of knowing how much is still on the surface. In general if it is not on the surface and has not been lost by accidental transfer, then it is either in or has passed through the SC. Also, if it is in the SC, then the chances are it will continue its journey over time—there’s no good reason for it to emerge again on the surface.
Stuff can be good. A scientifically perfect formulation for delivery is no good if all the other aspects of a cosmetic, i.e., looks, feel, greenness, cost, etc., are not right. But stuff can also be bad for delivering the active either because it is sitting uselessly in the way, e.g., 30% emulsifier, or has disappeared before it can do any good, i.e., the water required for an aqueous active evaporates. It is not a cosmetics crime to have stuff sitting uselessly in the way; in fact it may be inevitable due to the other needs of the cosmetic. What is a crime is to not have thought about the stuff in the first place.
Send e-mail to firstname.lastname@example.org.
- BM Magnusson, YG Anissimov, SE Cross and MS Roberts, Molecular size as the main determinant of solute maximum flux across the skin, Jof Investigative Dermatology 122 993–999 (2004)
- S Abbott and JW Wiechers, Formulating for efficacy, software, available at www.jwsolutionssoftware.com (Accessed Dec 4, 2012)
- S Abbott, An integrated approach to optimizing skin delivery of cosmetic and pharmaceutical actives, Intl J Cos Sci 34 217–222 (2012)
- CM Hansen, Hansen Solubility Parameters— A User’s Handbook, 2nd edn, CRC Press: Boca Raton (2007)
- B Geusens et al, Flexible nanosomes (SECosomes) enable efficient siRNA delivery in cultured primary skin cells and in the viable epidermis of ex vivo human skin, Advanced Functional Materials 20, 4077–4090 (2010)