June 13th, 2006

It is about time I tackled this rather monstrous subject because it is something that is important in what we do.
There are two foams of real interest to the barista – milk foam and crema. Before I go into detail on either I think it is best if we just started with a little basic foam theory.

So – why does causes a liquid to foam?

At some point a bubble has been introduced to the liquid. In a foam this bubble is made strong, because something has wrapped itself around it and made the bubble strong (increased its surface tension). In most food foams it is protein that wraps itself around the air. The reason that it does this is that part of the molecule is hydrophobic, and is repelled by water. For this reason it is looking for anything that isn’t water to face/bind too. Normally a protein chain is all coiled up, with the hydrophobic parts facing inwards, away from the water solution in which it is dissolved. It takes some sort of action, be it heat or – in the case of egg whites – an aggressive physical action like beating it.

So – the air bubble provides solace for the section of the protein repelled by water and what you get is the protein lining the air water interface. This is now a surface active agent, or in the parlance of foam – a surfactant.

What makes a foam burst?

If the surfactant is in water solution then a primary reason for failure in the foam is drainage. As gravity pulls the solution from between the bubbles, the surfactant is also drained away. This leaves the foam brittle and inelastic causing the bubbles to burst. This doesn’t explain why your souffle went horribly wrong. Here the enemy of the foam is fat.

Fat and water don’t mix. This is because one is polar (water) and one non-polar (lipids). The part of the protein that is hydrophobic is non-polar, hence is repelled by the water. If fat is introduced into the foam then it can give the non-polar part of the protein another choice, so it can either wrap around the air bubble of the fat. This is why fat destabilises foams, and your souffle went all to hell. (This is also why you need to make sure you never get any yolk in the whites if you want to beat them up).

A simple picture (cribbed from a text book and slightly changed.) By no means perfect but goes some way to explaining why surfactant bind to air bubbles:

Milk Foam:

In milk foam our surfactant is a whey protein, with the amazingly exciting name of beta-lactoglobulin (this is unlikely to be the only protein involved in the foaming process but it is the most important one). The reason most milks won’t foam is that the milk fat has broken down and the free glycerol from the tri-gylceride is interupting our foaming process.
Drainage plays a slightly different role for the barista. Drainage is linked to viscosity, to the thicker the liquid the slower the drain. This is why full fat milk is generally easier to work with. It will stay wet longer than skimmed that dries out pretty quick making latte art a little harder.

I will also speculate that you will get a better, more stable foam from your good technique (stretch only whilst milk is cool, churn/spin up to final temp) because you have formed your stable foams before the milk fat has reached body temp, at which point it turns from a fat to an oil, and is more likely to mess with the foaming process.

Any further questions on milk foam please leave a comment and I will answer (if I can!)


I don’t know why we make crema more complicated than it is. It is constantly horribly misdescribed. The most common description of it is “emulsified oils”. Well, I suppose, in a very very small way this is true. Espresso, as we know, is unique as a brewing process in the quantity of oil that is gets from the ground coffee. It does this by spending the energy of the brew pressure emulsifying the oil by creating such small droplets that they can be suspended in the solution without flocculating. Yet we know that oil causes foams to fail. The percentage of oil in espresso, as a total, is extremely low. Crema has nothing to do with the oils.

CO2 from the roasting process is liberated during the brewing process and the change in pressure causes it to come out of solution forming bubbles. So what is the surfactant here? It isn’t a protein (the roasting process has pretty much gotten rid of all the useful proteins here, as they’ve gone off to contribute to all sorts of reactions including everybody’s favourite friend – the maillard reactions). The acting surfactant is probably a derivative of a protein. It is a melanoidin. This is rather a broad term for molecules formed during the roasting process, and they also contribute to the brown colour. This is working with two other polysaccharides to create a stable foam. (The sugars role appears to influence foam volume and stability without actually being surfactants themselves).

Crema is essentially a foam of the coffee underneath – so the colour of your crema is a direct reflection of the colour of the coffee underneath. (This is a bit confusing because in a shot glass the liquid coffee from a 15 second shot looks just like the coffee from a 30 second shot – I assure you they are totally different colours, put a single drop on some white paper if you don’t believe me). So – the lighter the crema, the lighter the coffee, the less dissolved underneath.

There is speculation that the oil is a factor in how quickly the foam dissipates. The fact that arabica has nearly twice the lipid content of robusta goes a way to substantiating this. Along side this we also have drainage. A good shot has a substantially different viscosity to a poor one, which is why a quick shot’s crema will go very quickly. Again it is down to how much you have in solution, so colour of crema and stability should be linked.

Tiger flecking is something I don’t totally understand, and will be doing some research on soon. I currently speculate that it is particles of roast coffee suspended in the foam. Some microscopic photography is in order!

I hope to add more to this article as and when I can, and to answer questions that will inevitably come from it. Plus I will reword things I’ve probably explained really badly!

Questions in the comments please!

Comment Policy

There are no longer comments on new posts. If you'd like to respond you can find me on twitter.