Foam Making

It is trivially easy to make a foam - just mix air and liquid with some energy and bubbles will form. If these bubbles reach the surface with a liquid fraction ε in the 0.1-0.2 range then they are a kugelschaum ("kugel" means "sphere" and "schaum" means foam). These foams are not really considered in these apps. When ε <0.1 then we have a polyederschaum (polyhedral), the classic foam that is the central concern of Practical Foams. Although it is easy to create a foam, in most cases it is totally unstable. So the question of making foam is not so much about how to make them (which is trivial) but how to make them stable (which is not). In the AntiFoam section we will discuss the even more difficult question of how to make a stable foam unstable.

As noted in Basics, the energy required to create a foam is inversely proportional to the surface tension, γ. The low surface tension certainly helps, but if γ changes from 40 (a "bad" surfactant) to 20mN/m (a "very good" surfactant) it's only halved the energy needed, which isn't all that significant. Consider low γ as necessary (after all, pure water can't form a foam) but not sufficient. So what things are required?

1. Elasticity. The first reason surfactants help create foams is that the surface becomes elastic. This means that the bubbles can withstand being bumped, squeezed and deformed. A pure water surface has no such elasticity and the bubbles break quickly. It also means that those systems which produce more elasticity (see the Elasticity section) will, other things being equal, produce more stable foams. As discussed in the Rheology section, in general a wall which is both stiff and elastic provides a foam with a greater ability to resist a pushing force and therefore a higher yield stress. Smaller bubbles also give a higher yield stress
2. Disjoining pressure. The second reason that surfactants help create foam is that the liquid in the foam walls is naturally sucked out of the walls into the edges. This is nothing to do with drainage (as explained in Drainage, the walls contain an irrelevant fraction of the liquid), it is just simple capillarity. The capillary pressure will keep pulling liquid out unless a counter pressure ("disjoining pressure") acts against it. This can be produced by charges on the surfactant either side of the wall and/or by steric interactions between surfactant chains. These effects are discussed in DLVO, but because the charge effect operates over large distances (50nm) compared to the small distances (5nm) of steric effects, in general ionic surfactants are much better at creating stable foams.
3. Resistance to ripening. The Ostwald ripening effect means that small bubbles shrink and large ones grow. As the Ostwald section shows, this is partly controlled by the gas (CO2 falls apart quickly, air/N2 is slower and C2F6 much slower) but also by how good a barrier to gas diffusion the "wall" of surfactant at the surface provides.
4. Resistance to drainage. The more water around the foam the less risk (in general) of it becoming damaged. So a foam that drains quickly is more likely to become damaged. As we will see, to resist drainage you need high viscosity and small bubbles, though the surfactant wall has some effect on the drainage process with stiffer walls giving (usually) slower drainage.
5. Resistance to defects. If oil or a hydrophobic particle can penetrate the foam wall it can cause the wall (and therefore the foam) to break. Although there are plausible and simple theories (discussed in AntiFoams) of Entry, Bridging and Spreading coefficients they turn out to be of limited predictive value. Once again they are necessary but not sufficient. The key issue is the Entry Barrier. When this is high the foam is resistant to defects.

These principles are so easy, yet creating foams efficiently is surprisingly hard. Why? The key issue is timescales. If a surfactant is marvellously elastic and has a strong disjoing pressure and is a good gas barrier and has a high entry barrier it might (and usually does) fail to form a foam because it takes too long to reach the liquid/air interface and form its strong resistant domain so the foam has already collapsed. On the other hand, a surfactant that quickly reaches the surface to create an adequate elasticity and disjoing pressure will produce large volumes of foam - though the foam will collapse quickly, especially in the presence of oily impurities such as grease being washed from one's hands.

This leads us to the issue of Dynamic Surface Tensions. It would be wonderful to provide an app that fully described the complexities of DST and which therefore allowed you to produce a mixture with very rapid decrease of ST to give the fastest possible foaming behaviour. But my reading of the literature is that it is quicker to measure the DST behaviour using (most usually) a Maximum Bubble Pressure device (which creates bubbles over different timescales and therefore gives the surface tension at each of those timescales) than it is to attempt to describe the behaviour via theories. In particular, there are great debates about whether DST is limited by diffusion, by barrier entry and/or via the need to come out of a micelle before entering the interface. My reading of the excellent review by Eastoe¹ is that simple diffusion dominates and that the existence of micelles largely makes no difference because the timescale for a surfactant molecule to partition from the micelle is very fast even though the timescale for micelle formation/collapse is very slow. Of course one can find real cases of entry barriers and real cases of micelle-limited diffusion. But it is even more complicated. An extensive analysis from U. Sofia shows that there are 4 possible outcomes in systems containing micelles, two of which are indistinguishable (to the casual observer) from simple diffusion kinetics and two of which might be confused with barrier kinetics. Finally, distinguishing entry-barrier and micellar effects from the effects of small amounts of impurities in the surfactants is surprisingly difficult and for the practical formulator using commercial, unpurified surfactants there is little hope of understanding the subtleties of DST curves. The take-home message is "Don't formulate foams without measuring DST, but don't spend too much time theorising about why you get great results for some specific surfactant combination." I don't like writing such advice as I usually find that good models are the best way to avoid lots of lab experiments. However, the 2020 review paper, discussed below, contains a master-class on the relevant theory and concludes "The theory doesn't really help - just measure the DSTs" .

The harsh reality is that successful foaming agents tend to be mixtures, with all the complexities they induce. The ubiquitous SLES/CAPB (Sodium Laureth Sulfate/CocoAmidoPropyl Betaine) mixture happens to be made from two excellent fast foamers. The CAPB on its own produces a lot of stable foam, but is rather expensive. CAPB is especially good at creating a high entry barrier so is resistant to oils during the creation of foam. SLES on its own produces a lot of relatively unstable foam. A mix of the two provides a good balance of cost, foam and stability. However, adding a small % of lauric or myristic acid has a dramatic effect on foam stability. It increases elasticity but also slows down bubble growth (Ostwald ripening) dramatically, so the foam remains small. This has a big impact on the ability of water to drain from the foam - drainage speed goes as Diameter² - and the drier the foam the more easy (other things being equal) it is to break it apart. The long-chain acids on their own are useless as foaming agents (and as sodium salts are of modest foaming ability as common soap, easily wrecked by hard water). The combination of SLES/CAPB/Long-chainAcid is a potent mix for creating a foam with small bubbles and a long life-time. Indeed, a simple way to transform a hand-soap to a shaving foam is to add a few % of the long-chain acid.

But what about my surfactant system?

The rules for creating a good, stable foam (or, indeed, the rules for making sure that such a foam is not created) are simple and clear. So why is it so hard to create new foam formulations? The answer is that if you have the right set-up to measure all the basics: CMC, Γ_m, disjoining pressure v film thickness, interfacial elasticity and entry barrier then it's rather straightforward to make the best out of any set of surfactants and foam boosters you happen to want to use. The measurements can largely be automated so lots of formulation mixes can be screened quickly. One problem, as mentioned above, is timescales. Most measurements are made after comparatively long times so it needs extra time-dependent experiments to see if the appropriate parts of a surfactant blend will get to the surface fast enough to create a foam which then becomes stabilised as the slower components arrive to form a tougher surfactant layer. The other problem is that small additions of co-surfactants, foam boosters etc. can make a large difference, so it is necessary to carry out measurements on large numbers of samples. A robotic lab set up to do lots of high-throughput screening can do a lot of the hard work, but most of use don't have access to such a lab.

In the longer term, a theory that could predict the interfacial behaviour of mixtures of ingredients would make development of foam much more rational. But such a theory seems to be a long way off.

The view from 2020

I wrote this page in 2014-15 and had no reason to update it till 2020. To my surprise, what I wrote has stood the test of time. I've not changed any of the previous text, other than the DST sentence that refers the reader to here. But a masterly review², backed by a serious amount of experiment and theory, allows us to be a bit more specific. Again it is the team at Sofia, led by Prof Tcholakova, who have clarified the situation with five key points.

1. Although both non-ionics and ionics can produce excellent foaming, the non-ionics need to be above 95% of the full surface coverage of the interface (with a Gibbs Elasticity over 150 mN/m) before they will foam well - it's a sort of all or nothing. Ionics can start producing credible foam at 30% of their surface coverage (even with Gibbs Elasticity of just 50 mN/m), with a with a steady increase in production as you head to 100%. The reason is clear: steric stabilization of the foam interface works well, but only when there is near-full coverage; the interface can break easily if there is even a 5% gap in coverage. Charge stabilized ionics are much more forgiving.
2. The speed at which the surfactants generate the surface coverage is critical. Basically, if they get to the interface in a few 10s of ms, you'll easily get lots of good foam. This speed depends on concentration, CMC, surface mobility, salt concentratioin in no way that is readily extractable with 2020's theory/experiment (for some hints of the complexity, see DST-Choice, and read the master class on the theory within the paper, which concludes that it's not much help). This is sad in one way, but liberating in another. Just measure the dynamic surface tension at a 10ms timescale and tweak the formulation till you find a large reduction in surface tension. On a typical Maximum Bubble Pressure Tensiometer this 10ms timescale is measured at ~300ms (there's a fixed factor for any given MBPT device) because the real age of a 300ms bubble (it's expanding all the time) is only 10ms. The tradition from the Sofia school is to call the measured time (e.g. 300ms) τ_age and the scientific time (e.g. 10ms) τ_u for universal.
3. The foam at shorter timescales (in this paper, 10 shakes of their measuring cylinder) is not necessarily a reliable guide to the foaming after longer timescales (100 shakes). The faster-acting surfactants, not surprisingly, give more foam at short timescales, but the slower ones can catch up. As discussed in the next point, the foams tend to be self-limiting, so an initial advantage doesn't necessarily lead to a long-term advantage. Of course, for applications such as personal care, fast foaming is a requirement so this difference in performance is important. The point is that one has to be careful to distinguish different types of limiting factors.
4. This is only hinted at in the paper, but is linked to other Sofia work, with more published results promised. The amount and stability of foam gets limited by its own production method. To make more foam you generally need lots of smaller bubbles. These are created by whatever forces are able to trap air and squash, or shear bubbles so they get smaller. As the foam gets richer with smaller bubbles, it gets more viscous (depending on 1/Radius, see Foam Rheology), so at some point the forces are not large enough to deform the bubbles to something smaller. The effect depends somewhat on the rigidity of the interface and, therefore on the surfactant, but it's mostly dominated by the ability to create the fine foam in the first place, i.e. the interfacial stability and speed of reaching it. This is why plenty of surfactants can produce similar amounts of foam as long as they are present at sufficient concentration to meet the previous two requirements. Looking back at plenty of other foam papers I see that there is lots of confusion of cause and effect because like wasn't being compared to like. And because there is (rightly) a separate focus on the foam stability, for which we have the other apps on this site.
5. The team deliberately used "as is" surfactants because their impurities show up rather interestingly in the data. Measurements of % surface coverage come, of course, from adsorption isotherms CMC and Γ and these often show strange behaviour because of low levels of other components. This generally doesn't bother us, the surfactants are what they are, but they certainly complicate academic analyses when it's necessary to know, for example, whether you have 50% or 60% of surface coverage.

Foaming techniques

I had generally paid little attention to the different foaming techniques, but the remark in the previous section about foams being self-limiting made me realise that I've come across quite a few different methods.

1. Shaking cylinder. Put, say, 10ml of solution into a 130ml measuring cylinder and oscillate it, checking the volume of foam after a given number of shakes. If you get 90% trapped air then you are at 100ml, so finding whether you have 91, 92 ... gets tricky in a 130ml cylinder. My impression is that this sort of foam is relatively coarse, but I might be wrong
2. Ross-Miles. Put some test solution in the bottom of a tall cylinder. Now dropwise add more of the solution from the top. The drops smashing into the liquid below produce a foam. Measure the volume at the end of the addition, then, for stability, the volume after a few minutes. Amazingly, this is an industry standard test.
3. Blender. Just get a big blender and put in enough liquid to cover the blades. Whizz away and measure the volume by pouring the contents into a measuring cylinder. The fact that this can be done suggests that the foam is rather coarse, because a fine foam would be hard to pour..
4. Planetary mixer. Take you Kenwood Chef or equivalent with a wire whisk and watch what happens as the whisk turns on its axis while moving around on the other axis. A paper from the Sofia group shows a clear self-limiting effect once the foam gets thick enough to squash the surface waves which initially trapped the air, so this seems good for testing for the ability to create finer foams.
5. Sparging column. Blow air through a frit at the bottom of a column containing your foaming solution. You get some idea of the foamability and stability from the stable height of the foam, and/or you can measure the weight of foam coming over the top in a given time. More details are available on the Foam Fractionation page.
6. Micro-foam test. I once had to measure foamability using mg of surfactant and μl of solution. This was remarkably easy to do with a steady stream of air blowing through a very fine syringe needle into the solutions in micro-titre plates. It's a very good high throughput technique (which is why we developed it) to distinguish low, medium and high foamers and short, medium and long-life foam. It's crude but amazingly effective.
7. Compressed air foam. Mix your surfactant solution with some high-pressure air, let it travel down a pipe, expanding as it goes, and burst out onto, say, an oil storage vessel in flames. I once wrote an app for a fire-fighting project that required the theory of such a foam and needed some measurements to parameterise the theory. Unfortunately the live experiments on a full-sized test rig failed because the rig burned down during one of the tests...
8. Aerosol foams. This is a variant of the previous one, on a smaller scale. The propellant in a can (typically a hydrocarbon gas blend) is beautifully mixed into the surfactant mixture so creates a mass of fine bubbles when it suddenly expands. A typical example is a shaving foam which has to be fine in order to have the high viscosity and yield stress to stay on the face.
9. Hand rubbing. I know that foaming has no significance in terms of washing - the craving for it is psychological, not physics. So I'd never bothered to see how much foam one could create with imaginative hand rubbing. It's quite a lot, but in my view not worth the effort.
10. Shaving brush. I had never understood shaving brushes. They didn't produce an interesting amount of foam and just seemed a complicated way of spreading soap over my face. But then I'd never bothered to learn how to do it. If you whisk away onto a blob of wet soap on one's hand, nothing much seems to be happening. That's because all the foam is in the brush. Just squeeze the brush in any way, and out comes a mass of very fine, stable foam, perfect for placing on the face. I was very impressed.
11. Foaming net. Take a few cm of a fine net and rub it hard between your hands with the wet soap. As with the shaving brush, nothing much happens if you don't know what you are trying to do - I had to go to YouTube to find out. If you pull the net between your fingers, a large amount of foam emerges. Repeat this a few times and you get an awesome amount of fine, stable foam. The fine net is clearly good at breaking up larger bubbles into smaller ones. Why anyone bothers to spend their time creating this mass of foam bubbles is not a question I am qualified to answer.
12. Measurements of key parameters.
    • Obviously foam height, where appropriate, and the ratio of the total height to the amount of liquid in the bottom of the container, and how this changes over time.
    • A conductivity meter across a known gap, calibrated with the conductivity of the water used in the experiment, gives you a good idea of the volume fraction of air.
    • Put a large prism in contact with the foam and couple light into and out of it. A video shows a strong contrast between contact with water (white) and air (black) and it is then easy to use image analysis to measure the foam. Experiments have shown that the prism has a surprisingly small perturbation on the foam itself so the measurements are relevant. It's incredibly hard to get good image analysis from images of free foam because there's seldom reliable good contrast between walls and the rest.

Oil foams

It seems obvious that you can't make foams in oils. The surface tensions of oils are low and a surfactant can't make much difference and therefore the crucial elasticity stabilising effect cannot come into operation. This is generally true for simple hydrocarbon oils. To produce foams in these you need to use clever particulate tricks such as lyotropic phases of specific surfactants (such as mono-Myristylglycerate) or hydrophobised silicas (look up Binks in Google Scholar). But the real oil industry has massive problems with foams and the art/science of finding defoamers for each specific crude oil is a major challenge. Why do many crude oils foam?

The clearest scientific description of this comes from work by Callaghan and colleagues at BP³. They carefully extracted all the acidic components from a wide range of oils (these typically represented only 0.02% by weight) and found that the oil showed (a) no elasticity and (b) no foaming. If they added the extracts back to the no-foam oil then both elasticity and foaming returned. The acids were rather simple long(ish)-chain alkanoic acids such as dodecanoic. Although this paper did not record the surface tensions of the crude, other papers show typical values in the low 30mN/m but which can be reduced to the mid 20's by additions of simple surfactants or defoamers. This is not a huge decrease and, therefore, the elasticity effects cannot be large. However, in crude oils the pressures can be very high so the bubbling can be very violent when the crude reaches atmospheric pressure, so it doesn't need a very strong surfactancy effect to cause massive foaming.

Going back to the other type of foam stabilisation, crude oil is usually complicated by the presence of asphaltenes which can readily crystallise/cluster at the air/oil interface and provide foaming in that manner. And, as we will see, foam stability is greatly enhanced by high viscosity which many oils can readily supply. But nothing is simple: asphaltenes have been shown to be veryy modest surfactants that can produce foaming in toluene where they are (by definition) soluble.

Fire Fighting Foams

This is a huge subject. The only point raised here is that for oil/petrol fires the surfactant should not be good for emulsifying the oil with the water in the foam. The standard theory therefore states that the system needs a large "Spreading Coefficient" (see the Antifoam section) which in practice can only be achieved with fluorosurfactants. Such foams are astonishingly good at being jetted through huge flames to land nicely on the surface of the burning liquid (which, to the surprise of many, is "only" at its boiling point - not some super-high temperature) and put out the fire. For really robust foams adding a protein surfactant is a good idea - usually as part of a fast/slow mix of a normal fast surfactant to get the foam going and the slow protein which reaches the interface after a time and renders the whole thing remarkably solid. Alternatively some high MWt polymers can perform this function to create an AR-AFFF Alcohol Resistant-Aqueous Film Forming Foam which means one that works not only on non-polar fires but also on polar fires for which a conventional foam might be too compatible with the liquid.

However, with the move away from fluorosurfactants (seemingly inevitable, justifiable or not) my view is that it's necessary to focus on creating what I call LRLP foams, Low Radius and Low Permeability, created with standard surfactants. If you explore foam rheology, drainage, Ostwald ripening you will see that small-radius foams are stiffer and tougher. So you can gain foam lifetime via smaller bubbles. And with tricks like adding myristic acid, you can make a foam Low Permeability by making the interface stiffer. This helps reduce the rate at which warm vapours can move through the foam, reducing the risk of them re-igniting.