Crema: How does it form anyway?

Want to talk espresso but not sure which forum? If so, this is the right one.
User avatar

#1: Post by rpavlis »

When I got my used brass La Pavoni Europiccola made just before the change to the Millennium model I examined it very closely. I quickly decided that if I were to be able to make espresso that did not taste somewhat worse than dish water I would have to understand as much as I could about the process and always proceed in a very systematic manner. I described in a recent post some of my conclusions about the physics and physical chemistry of the process.

I was always curious about crema. I found two explanations of the subject, and one seemed not very feasible and the other seemed to have problems as well.

Explanation I: During the roasting process all sorts of carbon dioxide is formed. Somehow this CO2 is trapped inside the beans and when hot pressurised water is passed over it it escapes, and as the pressure falls as it leaves the coffee puck carbon dioxide bubbles form--crema. There are serious problems with explanation. CO2 is a weak acid. Roasted coffee beans are porous. The CO2 would have to be bound very strongly to the coffee for it not to escape! CO2 diffuses rapidly, several hundred millimetres per minute. Furthermore when liquids are supersaturated in gases, gas evolution tends to be slow--notice how slowly it escapes, for example, from club soda!

Explanation II: During the "pull" the pressure in the group is around 1000 kPa or 10 bar. At this pressure oxygen and nitrogen from the air become soluble enough to dissolve some in the hot water--when the pressure diminishes as the water moves through the coffee puck the pressure is lowered, O2 and N2 become less soluble and bubbles form. The problem is that the solubility seems too low to produce as much crema as is commonly observed, and also there is also the fact that gas evolution tends to be slow.

It seemed to me that BOTH of these explanations would result in MUCH more crema if I were to draw carbon dioxide into the group before the pull rather than air. I used a large coffee mug for a CO2 generating vessel. I filled it about 1/5 full of saturated sodium bicarbonate, and slowly added vinegar to it during the time that I pulled up the handle prior to making the pull. I even raised the lever half way and brought it down to purge the lower part of the group of air. With the bottom of the group in a CO2 atmosphere, I thus raised the handle all the way. I removed the large mug I had used to generate CO2, waited 25 seconds, and made the shot. THERE WAS NO MORE CREMA THAN WITH AIR!!!!!! The espresso was indistinguishable from normal, except I could taste the CO2 a bit, it was like having weakly carbonated espresso. It seems to me that this result means that BOTH of the above explanations are inadequate.

I restore old "top of the line" research microscopes as a pass time. I set up my Wild M40 inverted phase contrast microscope with a 10X bright contrast phase objective. I ground some beans with my "Rocky" grinder, suspended them in water, and examined them with this. The particle size was fairly uniform, with some of the usual tiny fragments that result from grinding things, some down to the size of bacteria. I prepared my first espresso for the day and put some of the crema on a slide between cover slips. It was a mass of bubbles that were quite uniform in size. (Similar in size to the size of the coffee particles.) There were bits of small particles stuck to some so that they looked like the foam that forms in flotation separation of materials from ores in mining operations.

It seems to me that the explanation for crema formation is very different from the two explanations I have seen before. When the water "floods" the coffee puck in the filter basket all of the air does not escape, leaving many spaces with air. These spaces are also relatively uniform in size. When the pressure rises in the coffee puck as the handle is pulled down these little gas voids are compressed to only about a tenth the size they had earlier from ten times as much pressure. They are small enough to pass through the spaces in the coffee puck, and when they emerge the gas expands about ten times and we have a bunch of little bubbles. Surface active things in the coffee prevent their collapse, and we have crema!

This explanation explains why the bubbles in crema are fairly uniform in size.

What is interesting is that this explanation is the simplest of all! It is the only one that seems to make sense to me. What do others think?

User avatar

#2: Post by bostonbuzz »

I like the visualization of your theory. Little compressed bubbles growing as they come out. It explains why pump machines GENERALLY have more crema- they have more pressure (again generally) to compress more bubbles to squeeze through the puck. As others have found, the more lever force increases crema.

However, I do think that the CO2 in the beans does play an important part. I may be misunderstanding you here, however. How do you explain how stale beans produce less crema, or that freshly roasted beans produce far more (not done outgassing). The fact that beans outgas over a period of days post-roast is not in question. Perhaps simple "air" is accounting for a percentage of the cream, but looking at the crema difference of stale and fresh roasts seems to say that that percentage is rather small. Somewhere the CO2 is coming out of the beans (becoming soluble?) and becoming crema.

Crema is complicated.
LMWDP #353

Baratza: skilled in the art of grinding
Sponsored by Baratza
User avatar
Supporter ❤

#3: Post by Peppersass »

Ah, an amateur coffee scientist -- a man after my own heart.

If you don't already have it, you need to get a copy of "Espresso Coffee, the Science of Quality", Second Edition, by Illy and Viani (with Liverani). This is the definitive scientific text on all things espresso, from the coffee plant, to processing of the raw bean, to roasting, to grinding, to storage and packing, to percolation, to taste in the cup, and more.

Your first explanation is closer to what Illy and Viani describe:
Dissolved gases, mainly CO2, quickly effervesce in the cup, and bubble up to build a layer of froth. This makes espresso a composite beverage where two distinct constituents are present: supernatant foam and underlying liquid.


Espresso foam -- also known by the Italian term of crema -- is by itself a biphasic system composed of gas globules framed within liquid films (called lamellae) constituited by a water solution of surfactant. These films tend to set in a configuration of two layers of surface-active molecules facing the gas, with water molecules between them. The high molecular force in the film allows its peculiar geometry: a bubble if isolated, or a honeycomb-like structure of many bubbles growing close together...
The error in your theory is that the CO2 formed during roasting isn't stored just in the pores of the roasted beans, from whence it escapes relatively quickly as you suggest, but is also stored in the intact coffee cells, from whence it takes much longer to escape. Furthermore, grinding does not release all of the CO2:
When broken down by grinding, coffee cells release their pyrolytic gas formed at roasting. This gas is mostly composed of CO2 and CO...

...However, the unfractured cells still present in the larger-size particles (over 50 um) keep their content of high pressure roasting gas, which will contribute to the formation of 'crema'... Inevitably, the gas will slowly escape during storage through microcracks, or through the natural porosity of the cell walls.
So, the answer to the subject question is this: CO2 that remains in the intact coffee cells is released by the pressurized hot water and combines with lipids and other insolubles extracted from the grinds to form the bubbles comprising crema foam.

The retained CO2 is why you have to let the beans rest for several days after roasting. A good bit of the CO2 needs to escape from the cells (i.e., the coffee needs to "degas"), or too much will be released when the hot water hits the puck and will interfere with extraction (i.e., the coffee will be under extracted.) You can see this effect when you drip-brew freshly ground coffee that hasn't been rested: you get a huge "bloom" of CO2 bubbles above the coffee. The bubbles get between the hot water and the coffee grinds, preventing contact and extraction. The result is under extracted coffee.

This also explains why stale coffee usually doesn't produce much, if any, crema. If the beans are allowed to sit for too long after roast, most of the CO2 escapes and you get no crema.

User avatar
rpavlis (original poster)

#4: Post by rpavlis (original poster) »

There are other simple experiments that can be done here. I do not have access to a vacuum line with gas chromatography equipment attached. One could use such a system to determine the composition of crema gases. It seems to me that a ridiculously large amount of CO2 would need to be stored in coffee beans to produce the observed effect. Furthermore, the fact that making espresso in a CO2 atmosphere does not increase quantity of crema dramatically seems to me highly significant.

If one examine dry roasted coffee beans under a microscope they are somewhat porous. It is hard to see how much CO2 could possibly physically trapped inside unless chemically bonded. One must also remember that gas bubbles are always trapped in powdered solids when liquids are poured over them. (I know this only too well from being a chemist and doing column chromatography. You must suspend the solid packing in solvent and add the slurry to the column rather than add the dry packing and pour solvent over it.)

Another simple experiment would be to place freshly roasted beans in a vacuum chamber for a few minutes and see if that resulted in loss of crema.

I suspect that stale beans have less surface active compounds in them so that the bubbles that would otherwise form tend to break. It is interesting that we have anti-foam materials that we deliberately put into distillation systems to prevent foam in laboratory practice. It would also be interesting to compare ground stale beans to fresh ones microscopically. Perhaps old beans are less brittle and are more "mashed" instead of cracked in coffee grinders. A problem is that there are too many variables.

I am sure that freshly roasted coffee contains things that react with water vapour and air. There are some interesting compounds in coffee. One is an interesting zwitterion, it is N methyl pyridine with a carboxylate anion in the beta position. Compounds like this may stabilise bubbles because they are "soap like".

There also can be multiple things going on here at once!


#5: Post by chang00 »

I think the stability of foam (bubbles in club soda vs crema), ie surfactant that was alluded in the earlier post was due to the galactomannan and arabinogalactan, described here in 1997 and from my recollection, also in the Illy book:

This was the substance that was previously described as "Hartmann" factor.

Recently I received the mass spectrum publication advertisement. The Wiley Registry of Mass Spectral Data, 9th Edition is $6875. The Flavors and Fragrances of Natural and Synthetic Compounds 2 is $2500. Anyone here still has access to GC/MS? :idea:


#6: Post by Anvan »

Mr. Pavlis, your thought experiments are interesting, and it's clear you enjoy musing on these topics. I encourage you to take Dick Green's recommendation literally and Amazon yourself a copy of Dr. Illy's seminal text. Armed with the established research, you may benefit particularly - besides no doubt enjoying the read.

User avatar
rpavlis (original poster)

#7: Post by rpavlis (original poster) »

One thing that is important here is that the tiny bubbles in crema must be largely water vapour in hot espresso because of the law of partial pressures. When crema cools it contracts dramatically, even without having the bubbles break up as the water vapour pressure falls. You can put crema on a microscope slide. As the water vapour condenses the bubbles shrink dramatically. I described before examining it with a microscope. (The room temperature was about 15C).

During roasting at high temperatures entropy considerations favour small molecules. On cooling these small molecules may react because the free energy will tend to favour larger molecules. Some of these reactions may release small molecules like water and carbon dioxide. In the case of roasted coffee it is apparent from the fact that it changes after roasting that there are many reactions that occur on cooling made possible by the entropy component of the free energy equation allowing the reaction to proceed. These reactions must continue to occur for a long time--that is why coffee "ages." Does it really lose mass as some imply?

The paper about the polysaccharides is interesting. The next time I have a chance to go to the University library I will try to find more than the abstract.

It seems almost certain that the gas in crema other than the water vapour is derived from the dissolution of high pressure air, from air entrapped in the ground coffee, and carbon dioxide. The question is which source provides the most. The whole thing is more complex than it seems at first.

Flair Espresso: handcrafted espresso. cafe-quality shots, anytime, anywhere
Sponsored by Flair Espresso
User avatar
Supporter ❤

#8: Post by Peppersass »

rpavlis wrote:If one examine dry roasted coffee beans under a microscope they are somewhat porous. It is hard to see how much CO2 could possibly physically trapped inside unless chemically bonded.

Another simple experiment would be to place freshly roasted beans in a vacuum chamber for a few minutes and see if that resulted in loss of crema.

I suspect that stale beans have less surface active compounds in them so that the bubbles that would otherwise form tend to break.
You're still ignoring the fact that a coffee bean is not a rock. It was once part of a living organism and contains intact cells, even after roasting. The CO2 produced during roasting is retained inside the cell walls. Porosity of the aggregation of cells has nothing to do with whether the CO2 is retained or not.

Your guess that stale beans have less surface active compounds, and that means bubbles won't form, seems entirely speculative. I've not seen any literature that talks about surface active compounds on coffee beans.

Again, get and read the Illy book before musing.

Placing freshly roasted beans in a vacuum chamber might pull CO2 out more quickly, but it would depends on factors like the permeability and/or strength of the cell walls. I don't know if such an experiment would be conclusive.

But one that you might try more easily would be to grind two shots of freshly roasted coffee. Pull one shot immediately and measure the crema. Put the other ground shot in a plastic bag, squeeze out the air, seal the bag and let it sit for a couple of days, then pull a shot with it and measure the amount of crema produced. My guess is that you'll get more crema with the first shot. You should also be able to see that the first shot contains larger bubbles, which are common in the crema produced by coffee that hasn't been rested.

I don't know if your attempt to pull a shot "in the presence" of CO2 is valid. My understanding from the Illy book is that the CO2 in the cells is released when contacted by the pressurized hot water, and that the CO2 dissolves in the water. When the liquid hits the cup, CO2 effervesces out of it and combines with some of the insolubles to make bubbles.

Gee, if roasted coffee doesn't contain a lot of CO2 that escapes over time, how do you explain the fact that virtually all roasters package coffee in bags with a one-way valve?