Monika Fekete investigates why milk bubbles form and how they impact on coffee quality and presentation.
Recently, I watched my coffee go cold in front of my eyes. It had been two months since I had enjoyed a proper coffee. I was looking forward to the experience as I sat at one of my favourite local cafés. However, no sooner had my coffee arrived, my newborn baby decided to test out his lung capacity and I found myself trying to calm him.
I had hoped this situation would be a one-off, but sadly it is not. My tiny son senses precisely when his mum is about to take a moment to enjoy her coffee, and duly demands attention. Consequently, I have had the opportunity to watch rosettas getting swallowed up in bubbles that slowly appear on top of my silky flat white. Interestingly, this wasn’t always the case. Sometimes many large bubbles emerged on the surface almost immediately, sometimes the micro-foam held together even after five minutes.
Curious to find out why some milky coffees develop bubbles while others don’t, I searched barista forums to check if others have made the same observation. The question is clearly out there, but I haven’t been able to find a commonly accepted explanation. Opinions vary greatly on the cause: Is it the coffee? Is it the milk? Maybe both?
So let’s investigate this bubbly mystery. Any scientific experiment is basically controlled observation. I needed to create an environment where I could observe and record what happens to a large number of milky coffees if I let them sit undisturbed for a few minutes.
I teamed up with an experienced barista trainer from Grinders Coffee (thanks for their assistance) who volunteered to pump out a series of highly consistent lattes, controlling the human element to his best ability (see main image). The factors we wanted to test for came up in online discussions: the degree and age of the roast, fat content and consistency of the milk, farming practice, and how dairy milk compares to alternatives, such as soy or almond milk.
Counting bubbles
We were interested in bubble formation over time, so I took a series of photos of each coffee, then processed them using an image analysis software. This way, I was able to count and measure circular features in a given size range (see image 1). The range was set so that micro-bubbles were not included in the calculation, only larger ones. The surface area covered by these larger bubbles, relative to the total surface area of the cup, became a measure of ‘bubbliness’.
The effect of coffee
To investigate the effect of the coffee first, I compared four identical lattes prepared with a lighter roast coffee and a darker roast coffee, both aged three weeks. Next, I compared another four sets of lattes made with a fresh batch (four days) and an aged batch (six weeks) of the same medium roast coffee. Identical amounts of homogenised, full cream steamed milk were added to all coffees.
The results are shown in image 3. Despite the large variation within datasets, a clear pattern emerged. ‘Bubbliness’ peaked around two to three minutes, then declined again as bubbles collapsed. The dark roast was slightly bubblier than the lighter roast, but the difference was not statistically significant for this sample set. The fresh coffee, however, was significantly bubblier than the aged one.
Both of these results can be explained by the release of CO2, where a fresh roast degasses intensely and releases bubbles into the milk, which then rise to the surface. An older roast still releases some CO2 upon grinding, but to a much lesser extent. It also makes sense that a darker roast would bubble more, as more CO2 is produced in a roast that had progressed further. This effect, nevertheless, was not strong enough to make a significant impact on milk bubbliness.
The effect of milk
We prepared a set of lattes made with a three-week-old medium roast coffee, replicating each condition four times. I compared full cream milk to skinny milk, homogenised milk to unhomogenised, standard to organic.
In terms of bubbliness, standard or organic farming did not make a difference. Skinny milk, however, was found to bubble around three times more intensely than full cream milk. Homogenised milk was also significantly more prone to bubble evolution than the unhomogenised version of the same milk.
So what difference does fat content make? Why is unhomogenised milk (where cream is collected at the top of a bottle) able to retain its microfoam structure better than its homogenised cousin (process in which fat droplets in milk are broken down)?
Read more:
• Water treatment around Australia
• How brew water temperature effects espresso extraction
• The affect of roasting and grinding on cell structure
When air is injected into milk, milk proteins (mainly caseins) surround the air bubbles, protecting them from merging with each other or bursting. This leads to the formation of microfoam. Microfoam degrades over time because gravity pulls their liquid coating down from the surface of the bubbles, which makes their walls become thinner. These weakened microbubbles first form larger bubbles then eventually collapse. So far, there is no difference between full cream and skim milk, their protein content is roughly the same.
The role of fat is more complex. Fat droplets can destabilise foam because milk proteins can become attracted to the lipid molecules instead of the air bubbles (Huppertz, 2014). This means that we could expect a more stable foam with low fat milk. On the other hand, higher fat content can help the initial formation of smaller bubbles (Oetjen et al., 2014, Munchow et al., 2015). My results are more in line with this second explanation.
Homogenisation breaks up fat globules into tiny droplets, which are stabilised by the same milk proteins that keep microbubbles separated. The larger total surface area of the small droplets absorbs more protein, leaving less to stick to air bubbles (Huppertz, 2014). This way, homogenised milk produces a less stable foam, which was reflected in the results.
Dairy alternatives
We also compared two brands of soy milk, a regular supermarket soy to a professional barista soy, as well as an almond milk. Soy milk was found to be significantly less bubbly than dairy milk, but there was no difference between the two brands. Almond milk bubbled somewhat more than soy. The difference is due to the nature of the proteins that help stabilise bubbles in each milk.
Are we closer to solving the mystery of milk bubbles? The mixture of coffee, milk and air is a physically and chemically complex system, where both coffee and milk make their impact on bubbliness. For a lasting microfoam, my advice is to avoid using very fresh roasts and maybe give unhomogenised milk a go.
For more information, visit www.coffeesciencelab.com.au