Milk science: proteins, fat and foam

What holds a bubble together

Widely documented

Cow's milk is mostly water, with lactose, fat and two families of protein, caseins and whey proteins, dispersed in it. When air is beaten or steamed into milk, proteins gather at the surface of each bubble and form a film that slows its collapse. This is textbook description, written without a source open.

The more finely the air is divided, the smaller the bubbles and the glossier the foam. Baristas call the result microfoam: bubbles too small to see one by one, with a texture often likened to wet paint.

What a review of dairy research says about fat

Sourced

A review of research on the foaming of milk was read here from its abstract only, and this section reports no more than the abstract says. It states that the presence of lipids, the fats, can be detrimental to both the formation and the stability of milk foams, particularly at a temperature where the milk fat is partly crystalline, for which it gives 20 degrees Celsius as an example.

The abstract also says that phospholipids, free fatty acids and partial glycerides strongly impair the foaming of milk. Free fatty acids are what milk fat releases as it breaks down; that it does so in milk that is old or has been roughly handled is general dairy knowledge and is not in the abstract. Together they are the usual reasoning behind a familiar piece of barista advice: when milk that used to froth suddenly will not, try a fresh carton before changing technique.

The abstract is about cow's milk and mentions neither steam wands nor plant-based drinks. What the body of the review says on either was not read.

Skimmed against whole

Tendency, not a guarantee

Baristas commonly find that skimmed milk makes a large volume of stiff, dry foam with little effort and that whole milk makes less foam with a richer, smoother feel. That fits the review's abstract, which says skim milk foams can be extremely stable, particularly when formed at about 40 to 50 degrees Celsius, and that fat can work against foam. No measurement comparing the two on a steam wand was read for this page.

Neither is the right milk. They suit different results, and the choice is as much about how the drink tastes as about how the foam behaves.

Working temperatures baristas use

Typical range

These are figures in common use behind a counter, not thresholds from a study. None was checked against a source for this page.

Practice, described as practice. Milks differ and the changes are gradual.

Why steamed milk tastes sweeter

Accounts differ

Warm milk is widely reported to taste sweeter than cold milk although steaming adds no sugar. The explanations usually offered are that sweetness is perceived more strongly in warm food and that heating changes the milk's smell in ways read as sweet. No study of either was read for this page.

One explanation has nothing behind it in what was read. It is sometimes said that steaming breaks lactose down into sweeter sugars. What is known is that lactose-free milk is made by adding an enzyme, lactase, which splits lactose into two sweeter sugars, and that such milk tastes sweeter for it. No source read for this page says that steaming does the same, and none was found that tested it.

Why plant drinks behave differently

Sourced

Plant-based drinks start with less to work with. A study that analysed a small set of shop-bought products found that those other than soy carried under two per cent protein, against 3.2 per cent in cow's milk. They were products bought in one country and not barista formulations, and the study did not test foaming.

A review of plant-drink foaming, read from its abstract, reports that these drinks vary far more widely than cow's milk in what their proteins are and in what state, and that the lipids and insoluble particles they contain can impair the stability of foam. It does not rank oat, soy, almond or any other drink, and neither does this page.

They also meet a problem cow's milk rarely has: separating or curdling in the cup. A study of one long-life soy drink in instant coffee found that whether it separated depended on temperature, on acidity and on how concentrated the coffee and the soy drink were, with more separation when hotter. A review of plant-protein creamers describes heat, coffee's acids, salts and polyphenols as promoting the clumping of plant proteins, and formulation as the means used to resist it. Applying the same reasoning to oat, almond or coconut drinks is an inference.

A plant drink is a suspension, built to stay one

Sourced

Cow's milk is an emulsion that an animal made. A plant drink is one a factory made, and a review of these products published in 2016 describes how. Plant material is broken down in water, the liquid is separated and then forced through a homogeniser, and the result is a suspension of fat droplets, protein, starch granules and fragments of the plant, most of them between 5 and 20 thousandths of a millimetre across. The review sorts the drinks by what they start from, cereals such as oats and rice, legumes such as soy, nuts such as almond and coconut, and seeds, and the starting material decides most of what follows: by its figures a dry soybean is about 40 per cent protein and an almond about 25, while an oat is mostly starch, with 11 to 15 per cent protein.

Staying mixed is the central difficulty. The review names three things that govern it, the size of the particles, whether a stable emulsion has formed, and how soluble the proteins are, and describes the means used: finer grinding and higher homogenising pressures to make the particles smaller, and added emulsifiers and thickening gums chosen according to the kind of drink and its acidity. It also notes that heat treatment changes the proteins and fats that hold the emulsion together. That is the background to the separation described above. A drink engineered to stay suspended in a carton at its own acidity is being asked to do something else when it is poured into hot, acidic coffee.

Oats bring a problem of their own, and it happens at the temperature of a milk jug. Starch makes up more than half of the grain, and the review gives the range over which oat starch swells and gels as about 45 to 74 degrees Celsius: heated as it is, an oat liquid thickens towards a paste. Manufacturers therefore break the starch down first with enzymes, and the review describes this as a step in making the drink. What a particular carton does on a steam wand depends on how completely it was done and on what else was added, and nothing read here measures that.

No study of steaming a plant drink on an espresso machine was opened, and none comparing the versions sold for baristas with the ordinary ones. What those versions contain is on their labels; what it does in the jug is the makers' claim and the experience of the people who use them.

What happens when coffee meets milk

Sourced

Milk does more to coffee than dilute it. Two things have been measured. The first is chemical. A review of how chlorogenic acid, coffee's main phenolic compound, interacts with food proteins reports that it binds to the caseins of milk by weak attractions, for the casein studied most closely chiefly the attraction between water-repelling parts of the two molecules, and without forming a permanent chemical bond. The studies behind that were made on purified proteins in the laboratory, not on a flat white, and the review does not say what the binding does to taste. It is often suggested as a reason milk softens the astringency of coffee; nothing read for this page tested that.

The second is about smell. In a device that imitates swallowing and breathing out, raising the milk fat in a coffee-flavoured drink significantly reduced the release of the aroma compounds that dissolve most readily in fat, while the non-fat solids of milk had a smaller effect. In a separate study three trained panellists drank black coffee and coffee with whole milk, at room temperature, while the air leaving their noses was collected: most odorants came through in similar amounts, and three were significantly lower with milk. Both studies were made by authors working for food companies, and the first used drinks at room temperature too.

Together they give a mechanism for something every drinker knows, that coffee with milk smells and tastes of less coffee than its dilution alone would explain. Fat holds back part of the aroma, selectively, and protein binds part of the phenolic material. One of the two studies went on to a tasting: when the compounds that milk had held back were added again until the breath matched black coffee, a company panel of 31 rated the milk coffee as smelling more of coffee. How large either effect is in a hot drink, and whether skimmed and whole milk differ in the cup as the first study implies they should, has not been tested by tasting in anything read.

A model of how much water steaming adds

CoffeeHQ explanation

A steam wand heats milk by condensing steam into it, so steaming adds water. This is a model of how much, worked from textbook constants. No measurement of it was read, and a real jug will differ for the reasons in the last row.

Arithmetic on textbook values for the heat capacity of milk and the heat of condensation of water. It says a tenth is the right order of size, not that any particular machine adds that amount.

From mechanism to a first thing to try

CoffeeHQ explanation

Large, soapy bubbles

Usually because: Usually air added late or roughly, so it was never divided finely.

Change first: Add air earlier and more gently, then keep the milk turning.

Milk too bubbly

No foam from milk that used to work

Usually because: Often the milk: age, or fat that has begun to break down.

Change first: Try a fresh carton before changing technique.

Milk not foaming

A plant drink splits in the cup

Usually because: Heat and the coffee's acidity acting on its proteins, as far as the soy study goes.

Change first: Let the coffee cool a little before combining them.

Plant milk splitting

What this page leaves to the technique pages

CoffeeHQ explanation

It gives no temperature as correct for steaming and no figure for when milk proteins change. It does not rank dairy milks or plant drinks for foam. It does not say why steamed milk tastes sweeter, only which explanation it found no support for. It makes no statement about whether a drink that has separated is safe to drink. And it does not teach the movements of steaming, which the technique pages do.

What to do next

• Steaming and frothing milk — The practical guide: wand, jug, frothing without a wand, and plant drinks at the wand.
• Milk frothing troubleshooter — Starts from the fault and the milk and gives one change to try first.
• Steaming plant milk — What baristas do differently for oat, soy and the rest.
• Milk types compared — How each common dairy and plant drink behaves in coffee, one page each.