Dalgona Coffee Explained

April 22nd, 2020

What keeps foams stable?

Foams are often described as inherently unstable systems. In my entry on mousse cake, I discussed how foams are defined as gas bubbles dispersed in a liquid phase. Naturally these two phases have very different densities and will tend to separate. Driven by liquid drainage and the minimization of surface area, bubbles tend to combine with each other, destroying or inhibiting the formation of a packed foam structure. Thus, the foams we encounter in daily life almost always need something else besides just the two phases; they need an additive that provides a supplementary stabilization mechanism that reduces the tendency for bubbles to combine. This stabilization can come in several forms; the most common for people to have at least heard about are molecules that are surface-active agents or “surfactants”, which is what soap is primarily composed of. As the namesake implies, surfactant molecules tend to locate themselves at surfaces like the interfaces between gas bubbles and a liquid in a foam. By doing so they lower the surface tension between the two phases, reducing the driving force for bubbles to combine. Though surfactants can significantly increase the lifetime of foams, liquid drainage will still occur, followed by bubble coalescence and the collapse of the foam structure.

An alternative method that has led to the creation of “ultrastable” foams is to saturate the interfaces between the gas bubbles and liquid with colloidal (microscopic) particles. Unlike surfactant molecules which will dynamically pop on and off an interface, solid particles are largely stuck in place, and form a physical barrier to coalescence that results in the “ultrastable” moniker for these foams. Even if liquid drainage occurs, these foams are extremely resistant to coalescence and collapse. The usage of these materials that I’m most familiar with is as yield-stress fluids for direct-write 3D printing inks. The particles stabilize the foam and after a shape has been 3D printed, the structure is heated, fusing the particles into a permanent and porous structure.

For both of these stabilization methods, the concentration of the stabilizer and aspects of the continuous phase are still important. Too little surfactant or particles at an interface will be ineffective at keeping bubbles separate. Regarding the continuous phase, I have been mentioning the concept of liquid drainage leading to bubble coalescence. A highly viscous continuous phase dramatically reduces the speed of drainage in the thin liquid films between bubbles, preventing the bubbles from directly contacting each other in the first place. This seemingly is key in the formation of Dalgona coffee which is why a significant amount of sugar is added. The sugar-water mixture composes the continuous liquid phase of the coffee foam and is very viscous.

So what is responsible for the foaming in coffee?

Not just in Dalgona coffee, “crema” coffee foam is a key part of the experience of drinking an espresso. Several naturally occurring substances have been identified in coffee that may act as surfactants including melanoidin protein complexes, and various aspects of crema formation have been considered in detail. Coffee of course also contains particles of ground coffee beans; these compose the dried stains of coffee and are responsible for the well-known “coffee-ring” effect . Both stabilization strategies that I’ve mentioned seem to possible play a role. With the experiments that I’ll describe next, I attempted to investigate the stabilization of coffee foam, and why there seems to be a difference in the foaming performance of instant and brewed coffee.

Experiments

Dalgona coffee. For the Dalgona coffee I followed a recipe that is widely cited: 2 Tbsp instant coffee powder, 2 Tbsp sugar, and 2 Tbsp of hot water (approximately 80 °C) in a bowl. I used Lavazza brand Intenso instant coffee powder. Since I’m fairly lazy, rather than using a whisk as is generally called for, I used an electric hand mixer. After about 5 minutes on medium speed, I obtained the delicious result that was quite stable over the course of an hour as you can see in the images.

The Dalgona coffee foam was quite puzzling in its rheological behavior. In the heuristic test of peak formation, this foam seemed to be somewhere between a soft and stiff peak, hinting at a yield stress around 3 to 5 Pascals. However, the post-yield viscosity was noticeably higher than in egg white foams that I’ve made, likely due to the higher concentration of sugar in the liquid phase. In addition to changing the perceived thickness of the foam, the higher viscosity seems to also give this foam a relatively long “relaxation time”, a concept that I’ll discuss more another time. For now, the relaxation time can be thought of as the time it takes for a material to react to a stress or deformation. For the Dalgona coffee foam, what a long relaxation time means is that though I was able to form peaks on the tip of my hand mixer (suggesting the presence of a yield stress), those peaks would noticeably droop and fall apart after tens of seconds. Since this foam is able to support the chocolate particles I sprinkled on top and can hold very small surface peaks after one hour, it’s likely that this foam does have a minute yield stress around 1 Pa or less. However, we also see a small amount of coarsening of the foam on the surface, bubbles have combined and become larger after an hour.

Looking only at the type and formation of peaks would have overestimated the yield stress. This teaches us that one of the key features of identifying a yield-stress fluid is to be able to observe it over a long period of time. This can be difficult or impossible to accomplish with less stable foams, as it isn’t always clear if any observed deformation is due to the material flowing or degrading over time.

Brewed coffee. For regular brewed coffee, I used my espresso machine to make around 100 mL of espresso from Lavazza brand Qualita Oro whole beans. From this espresso I took approximately 38 grams of liquid (equivalent weight to 2 Tbsp of the instant coffee powder + 2 Tbsp water) and added 2 Tbsp sugar. Following the same foaming procedure I was able to obtain a light-colored foam that was a bit less stable than the Dalgona, and which I do not believe is a yield-stress fluid. Looking at the surface of the foam, the coarsening is present here to a greater degree compared to the Dalgona, with the foam having pulled away from the edges of the glass in places after one hour. The surface is totally flat and has essentially no features that have held their shape. Chocolate particles sprinkled on top immediately begin to fall through the foam.

After making this material, it struck me that the observed difference in foaming ability for instant and brewed coffee may not be due to any inherent difference in the nature of the surfactants or particles that are present, but rather may just be a concentration issue. As I said, a high enough concentration of surfactant or particles are necessary to saturate the bubble interfaces. With instant coffee you are able to independently control the relative masses of coffee and water, and thus the concentration of the stabilizer can be very high. With brewed coffee this isn’t the case; to extract more out of the ground coffee, you also have to force more water through them. With this in mind, I attempted an experiment to concentrate brewed coffee to see if I could improve the foaming performance and stability.

Brewed coffee concentrated. To concentrate the stabilizers in brewed coffee, I first prepared one cup (approximately 235 mL) of espresso from my Lavazza beans. I simmered the espresso over low heat to evaporate off the water and hopefully leave any of the stabilizers unaffected. I stopped simmering once I had around the 38 grams of liquid that I needed (~6x reduction in volume), adding sugar and foaming as usual. The results were quite similar to the unconcentrated brewed coffee. The surface was still quite flat but I was able to support chocolate particles, suggesting a slightly increased yield stress and therefore a higher volume fraction of air that could be incorporated. However, the coarsening I observed was much more noticeable in this case, with the foam totally collapsing at the edges of the glass.

Having not diligently performed repeat experiments, it appears that significantly concentrating the coffee evaporating with heat does not significantly affect the foamability or stability of the material. The Dalgona (instant coffee) foam still seems to give the best results in terms of the stiffness of the initial foam and stability over time.

While it is possible that an even more intense concentrating of the brewed coffee or concentration by a different method might be yield different results, these experiments suggest that in fact there IS some inherent difference between the makeup of brewed and instant coffee that affects foam stability. With this in mind, I performed one more experiment to attempt to decouple the effects of the natural surfactants and particles as much as possible.

Filtered instant coffee. In an ideal situation, to independently test the performance of the surfactants and particles, I would be able to dissolve the instant coffee in water and then completely separate any particles from the liquid which would hopefully contain the vast majority of the surfactants. In a lab I would probably attempt this using a high-speed centrifuge to have the particles settle at the bottom of a container, then carefully remove the liquid via pipette. At home, the best I can do is to take the dissolved instant coffee and pass as much of it through a coffee filter as possible. I poured the 2 Tbsp instant coffee, 2 Tbsp water mixture over a coffee filter and left it for about a half an hour, covered in plastic wrap to reduce evaporation. At some point the filter clogged and I added another 2 Tbsp of water to try and pass as much of the original liquid through as possible. After another half hour I was left with a significant amount of wet sediment in the coffee filter and 3 ~ 4 Tbsp of filtered coffee. I obviously was not able to remove the smallest coffee particles by this method, but I was surprised at how much sediment came out of the instant coffee. Taking the filtered coffee (without most of the particles), I again simmered it down to approximately 38 grams, added sugar, and foamed it.

This time, I was able to again obtain deceptive medium-soft peaks but they were quite short-lived. After an hour this foam had completely disappeared making it the least stable of all the materials I tested. Since it is extremely unlikely that the surfactants were removed by the coffee filter, this experiment suggests that the particles that I removed are in fact key to the stability of Dalgona coffee foam. That isn’t to say that the surfactants in coffee don’t play a role, but they may not be powerful or concentrated enough on their own to maintain the stability of foams with a high volume-fraction of air like the Dalgona coffee foam. From a design perspective, if we wanted to start changing things about this foam like reducing the amount of sugar in the continuous phase or making stiffer, more airy foams, these experiments suggest that the natural surfactant concentration is not a tuning knob that we can rely on for coffee foam stability and that something else may be needed.

Particles in instant coffee

There are numerous other experiments that we could do to try and confirm our preliminary conclusion that the presence of particles in instant coffee is key to the stability of Dalgona foam, but it’s probably more important to consider a reason why it might be the case that the particles in instant coffee are different from those in brewed coffee and how that might affect stability. Certainly, allowing instant and brewed coffees to sit for a while results in dramatically different amounts of sediment gathering, in favor of instant coffee. Reading about instant coffee it seems that the most common production method is via spray drying brewed coffee. I have a little bit of knowledge regarding spray drying as it relates to my work with pharmaceutical manufacturing. In my field, spray drying of pharmaceuticals is notable for producing relatively uniform and spherical particles, something which is also noted in the Wikipedia page for instant coffee. If this is the case, I would imagine that when instant coffee is dissolved in water, the particles are much more spherical and uniform than the particles in brewed coffee that are produced directly from grinding. This could be a possible explanation for the improved stability of instant coffee, since irregularly shaped and sized particles often pack together more tightly than uniform spherical particles. A tighter packing density would require much more particles to span the area of interface and/or the liquid volume between bubbles, thus making stabilization more difficult for the brewed coffee. This is simply a hypothesis I have that would require many more experiments and characterization techniques to confirm, but I think it is plausible for explaining why instant coffee is better for foaming and why concentrating the brewed coffee did not make a significant difference. Some time soon I hope to revisit Dalgona coffee and improve it based on some of the principles we’ve explored here. Hopefully this gives you something interesting to think about while you stay safe and enjoy some coffee.