William Lee

Bubble Nucleation, Questions and Answers


Stout beer

These are some of the questions I've been asked about the background to and implications of a paper I and my colleagues Scott McKechnie and Michael Devereux submitted to Physical Review E (now in press) and uploaded to the arXiv preprint server.

Who did what in the paper?
I came up with the initial idea for the project. Scott and I worked on developing the mathematical model. Michael collected the images.

What advantages does nitrogen give to stout beers?

  • Less acidic taste: carbon dioxide increases the acidity of the beer, nitrogen does not.
  • Creamy mouthfeel: the bubbles in the head are very small which in turn is due to the low solubility of nitrogen in the beer.
  • Waves of sinking bubbles at the side of the glass: small bubbles are more easily carried along by currents in the glass.
  • Long lasting head: the low solubility of nitrogen slows the coarsening of the foam in the head.

Can you explain how the project came about?
The problem of modelling foaming in stout beers first came up at the 70th European Study Group with Industry. I was part of a team of mathematicians that looked at modelling the foaming of stout beer by ultrasound. Following the study group I remained curious about foaming in stouts and I thought it would be an interesting project for an intern to understand why the foaming mechanism in carbonated beers and champagnes did not work in stouts. I was surprised to find it did work, albeit slowly, and even more surprised when a rough estimate suggested it could potentially be useful.

I thought bubbles formed on imperfections in the glass surface. Is this not true?
This was the accepted explanation for bubble formation until recently. But when researchers looked at bubble formation sites under a microscope they found that in most cases the bubbles were nucleated by cellulose fibres. Scratches in glass can nucleate bubbles, but cellulose fibres are both more numerous and more efficient nucleation sites.

How much does the current widget add to the cost of manufacturing canned stouts?
So far as I know, the cost is relatively small: only a few euro cents per can. But there is also the problem of removing oxygen from the widget. Any traces of atmospheric oxygen would affect the flavour of the beer, and removing the oxygen is a time consuming process.

Where do the bubbles come from when draught stout is served in a pub?
In a pub, the dispensing mechanism forces the stout at high pressure through a plate with tiny holes in it. The turbulence generated by during this creates the tiny bubbles that in canned stouts are created by the widget. I don't understand this process as well as I would like and hope to investigate it in the future.

Is there an intuitive explanation for why a mixture of nitrogen and carbon dioxide gasses nucleate bubbles on cellulose much more slowly than carbon dioxide?
The key to understanding this is the very low solubility of nitrogen (1/50th that of carbon dioxide). This means that the concentration of nitrogen in solution is 1/50 that of carbon dioxide at the same pressure. Rates of diffusion (e.g. into a growing bubble) are proportional to concentration. So the rate of bubble growth is much slower than expected. The presence of some carbon dioxide in stouts speeds up the process, which is why bubble growth is 15 times slower and not 50 times slower.

What sort of geometry for the 2.9 cm cellulose square do you have in mind for the widget replacement? Is it something that would line the can?
I would suggest a band round the top of the can or the neck of a bottle (on the inside) so that the stout flowed past it as it was poured out. I think the ideal geometry would have the fibres perpendicular to the surface in a regular array. Our future work in this area may focus on determining an optimal geometry for the fibres.

Where do the cellulose fibres that nucleate bubbles in carbonated beers, sparkling wines and champagnes come from?
The cellulose fibres will either have been shed from the cloth used to wipe the glass dry or will have fallen out of the air.

Isn't it more likely that the fibres nucleating the bubbles are left over from the brewing or fermentation process?
In order to nucleate bubbles fibres must contain a gas pocket. It is likely that any gas pockets in fibres left over from the brewing process would have become flooded over time. Also if you look at a glass of champagne or carbonated beer you will see that most bubbles form on the walls of the glass with very few, if any, bubbles forming inside the beer. This is hard to explain if the nucleation sites are inside the beer.

What are the broader implications of the work, beyond stout beers?

  • One interesting discovery we made is it seems to be easier to study bubble formation in cellulose fibres in stouts than it is in carbonated liquids. The bubbles grow more slowly making them easier to observe, and when they reach the surface of the liquid they do not rupture (coating the microscope lens with microdroplets). So it may be that this study, which owes to much to previous work on champagne and other carbonated liquids, may be able to give something back.
  • Another possible idea is developing a microwave milk frothing device. Cappuccinos also require very small bubbles and it might be possible to devise a vessel which nucleated tiny bubbles on its walls as the microwaves heat the milk to its boiling point.

I'd like to see cellulose nucleating bubbles in stout for myself, how do I do it?
First you need a can of stout. You should be able to hear the widget rattling around in the top of the can. The problem is that if you open the can normally, the widget will create lots of bubbles which will scavenge all the dissolved gasses from the stout leaving none for the cellulose. So take some bluetack and a pushpin. Put the bluetack on the top of the can away from the tab of the ring-pull. Push in the pin through the blue tack and through the metal of the can (be careful and perhaps wear eye protection). Wobble the pin from side to side until you can hear gas escaping. Let the gas escape slowly, it should take a minute or two for the can to reach atmospheric pressure (check this by squeezing the sides of the can). At this point open the can as normal and pour slowly into a glass without splashing. The beer should look like flat coca-cola. However if you shine a powerful torch up from the under the base of the glass you may see trains of bubbles from (invisible) cellulose fibres rising very slowly. Pour some stout over a coffee filter and you should see bubble formation, especially under a microscope.

Who are the other researchers involved in this area?

  • Gerard Liger-Belair and coworkers provided the foundations for this work: we used their model of bubble formation as the starting point for our work.
  • Michael Chappell and coworkers also developed a model of bubble formation in carbonated liquids.
  • Andrew Alexander was involved in proving that bubbles in stouts do sink.

University of Portsmouth
Department of Mathematics


University of