It has long been known that nucleation sites are essential to give carbonated drinks their fizz. Spontaneous bubble formation at a smooth surface only occurs in liquids that are supersaturated with dissolved gas. Without nucleation, a glass of champagne would seem lifeless and still, and a pint of lager poured from a can or bottle wouldn't form much of a head. In practice this isn't a problem, as preparing a glass that it is free from the impurities that generate bubbles — such as cellulose fibres from the environment — is actually quite difficult.

But for stout beer, which is infused with a mixture of carbon dioxide and nitrogen gas, this isn't enough. To try to figure out why, William Lee and colleagues have extended a mathematical model that successfully describes the formation of bubbles in a purely carbonated liquid to one that describes a liquid containing a mixture of dissolved gases (W. T. Lee et al., http://arxiv.org/abs/1103.0508; 2011).

Credit: © 2011 GUINNESS & CO

There are many reasons for using nitrogen in place of some of the carbon dioxide in stout. The lower concentration of carbon dioxide lowers the beer's acidity. And bubbles of nitrogen tend to be smaller and more numerous than those of carbon dioxide, giving stout its characteristic creamy texture. This doesn't present any problems for draught stout served in a pub, which is forced through a perforated metal plate that agitates the beer to produce the requisite bubbles.

But when delivered from a can or bottle, the agitation is much less vigorous, producing fewer bubbles and a less fulsome head. To compensate for this, many canned stouts use a hollow ball — commonly referred to as a widget — containing pressurized nitrogen that rushes out when that can is opened, to break up into hundreds of millions of tiny bubbles in the resulting turbulent flow.

Lee and colleagues wondered whether nitrogen bubbles simply don't nucleate by the same process that carbon dioxide bubbles do, and whether it might be possible to promote their formation by some other mechanism. The model they subsequently developed suggests that nitrogen bubbles can in fact form by nucleation, but they do so at a rate that is much slower than carbon dioxide.

As a test, they immersed a cellulose fibre drawn from a coffee filter in a glass of canned stout and observed it under a microscope. They found that bubbles did indeed grow slowly, but regularly, from the fibre. This prompted them to speculate that coating the inside of a can with such fibres might produce enough bubbles quickly enough to remove the need for pressurized widgets. But whether that is economically viable, they leave for a more sober head to decide.