What temperature should you steam milk to?

Between 50 C and 60 C, which is about 122 F to 140 F. That range comes from a 2022 International Dairy Journal study by Klimanova and colleagues, which steamed milk from 30 C to 80 C and measured foam stability, viscosity and bubble distribution; pasteurised whole milk gave the most stable foam. Past roughly 70 C the milk tastes scalded and the foam turns stiff and dry, and neither can be reversed. Without a thermometer, stop when the jug first becomes uncomfortable to hold, and stop a little earlier than feels right because the milk keeps heating for a moment after the steam goes off.

How do you froth milk without a frother?

Fill a jar no more than a third full with cold milk, seal it, shake hard for about 45 seconds, then take the lid off and microwave it for 30 to 45 seconds. Shake cold and heat afterwards: a sealed jar of hot milk builds pressure, and foam made in hot milk collapses almost immediately. A French press works the same way and gives a finer foam, because plunging warm milk forces the bubbles through the mesh. Neither method produces microfoam, so the result sits on top of a drink rather than pouring through it, and latte art is out of reach.

Why is my steamed milk full of big bubbles?

You spent too long adding air, or the steam tip lifted out of the milk while you were doing it. Air should go in during the first two or three seconds only, as a quiet regular hiss with the tip just under the surface. After that, sink the tip deeper and let the milk spin in a whirlpool, which folds that air through the milk and breaks the larger bubbles down. Loud tearing or spitting means the tip is out of the milk and you are making bubbles rather than foam.

Can you steam oat milk?

Yes, but use a barista edition. Ordinary oat milk has little of the protein needed to hold a bubble open, and espresso is acidic enough to make it separate into grains. Barista versions fix both: Oatly's US Barista Edition adds rapeseed oil, which gives the foam the body dairy fat would, and dipotassium phosphate, an acidity regulator that stops the espresso splitting it. Soy foams better than most other plant milks because it carries real protein, while almond is the hardest of the common ones since it is short on protein and fat alike.

Can you re-steam milk?

You can, but it will not foam properly the second time. Steaming uses up much of the protein structure that holds the bubbles open, so re-steamed milk gives a thin foam and picks up a slightly cooked taste. Steam only what you are about to pour. Leftover steamed milk is perfectly fine to drink cold or to cook with, it just will not foam again.

Is whole milk or skimmed milk better for foam?

It depends which drink you are making. Skimmed milk foams more readily and gives more volume, because fat interferes with foam formation, but the result is dry and stiff and pours like meringue. Whole milk gives less volume and a much finer, glossier texture that pours as one liquid, which is what a latte or a flat white needs. Use skimmed for a tall cappuccino and whole for everything else.

How to Steam Milk (and How to Froth It Without a Machine)

What steaming milk actually does

Two separate things happen when you steam milk, and people who struggle with it are usually doing both at once. The steam pushes air into the milk, and the steam heats the milk. Foam comes from the first. Sweetness and texture come from the second.

The foam itself is built by protein. Milk proteins gather around each air bubble and hold it open, which is why a milk with more protein foams more readily. Fat does the opposite job: it softens the foam and gives it the heavy, glossy feel that makes a flat white taste rich rather than airy.

What you are aiming for has a name. Microfoam is milk where the bubbles are too small to pick out individually, so the surface looks like wet paint and the whole jug moves as one liquid when you swirl it. Stiff white peaks sitting on top of thin milk underneath are not microfoam. That is two separate things in one jug, and it will pour in two separate stages.

The two phases, in order

Purge the wand first. A second of steam clears the condensed water sitting in the tip, which would otherwise go straight into your milk and thin it.

PHASE ONE, AIR. Put the tip just under the surface and open the steam fully. You want a quiet, regular hissing for about two or three seconds while the level in the jug rises slightly. Loud tearing sounds mean the tip has come out of the milk and you are making bubbles rather than foam.

PHASE TWO, TEXTURE. Sink the tip deeper and angle it so the milk turns in a whirlpool. Nothing more should go in from here. The spin folds the air you already added through the rest of the milk and breaks the bigger bubbles down, and it keeps going until you stop it.

Getting these the wrong way round is the single most common mistake. Air added late sits on top and never integrates, because there is no spinning left to fold it in. Add your air in the first few seconds, while the milk is still cold, then spend the rest of the time texturing.

Air first, then texture, and never the other way round. A few seconds of quiet hissing near the surface, then sink the tip and let it spin until the jug is too hot to hold comfortably. Milk that has been overheated cannot be rescued, so stop earlier than feels right while you are learning.

The temperature, and what the research actually found

Almost every guide tells you 140 F, or 60 C, and almost none of them say where the number came from. There is a study behind it. Klimanova and colleagues, publishing in the International Dairy Journal in 2022, steamed milk across the full range from 30 C to 80 C on an espresso machine and measured what came out: foam stability, viscosity and how the bubble sizes were distributed.

Their optimal range was 50 C to 60 C, so roughly 122 F to 140 F. Note that the popular number sits at the top of it rather than in the middle. The same work found that pasteurised whole milk produced the more stable foam, with bubbles better distributed at 60 C, which makes the pasteurised against UHT choice a real variable rather than a preference.

Past about 70 C the milk starts tasting scalded, and that flavour cannot be reversed. Proteins that have been taken too far lose the elasticity that was holding the bubbles open, so an overheated jug gives you stiff, dry foam and a flat, cooked taste at the same time.

You do not need a thermometer to work in this range, although one is cheap and removes the argument. The usable heuristic is that the jug becomes uncomfortable to keep your hand on at somewhere close to 60 C. Stop when it becomes uncomfortable rather than when you cannot hold it at all, and stop earlier than feels right while you are learning, because a jug carries on heating for a moment after the steam goes off.

Which milk foams best

Whole dairy milk is the default for a reason: enough protein to build the foam and enough fat to keep it soft. Skimmed milk will foam more readily and give you more volume, but the result is drier and stiffer, and it pours like meringue rather than paint. If you want a tall cappuccino, skimmed is genuinely easier. For a flat white or a latte, whole milk is doing something skimmed cannot.

Plant milks are a different problem, and the barista versions exist to solve it rather than to charge you more. Two things go wrong with ordinary oat milk in coffee. It has little of the protein that would hold a bubble open, and espresso is acidic enough to make it separate into the grainy mess people describe as splitting.

You can read the fix on the carton. Oatly's US Barista Edition lists an oat base, then low erucic acid rapeseed oil, then dipotassium phosphate, and declares 3 g of protein against 7 g of fat per cup. The oil is there to do the job dairy fat does, giving the foam body. The dipotassium phosphate is an acidity regulator, buffering the milk against the espresso so it stays smooth instead of curdling. Any barista edition worth buying is doing some version of the same two things.

Soy foams better than most plant milks because it genuinely carries protein. Almond is the hardest of the common ones, since it is short on both protein and fat, so expect a thin foam that collapses quickly even in a barista version.

How to froth milk without a frother

None of what follows makes microfoam. Be clear about that before you start, because the internet is not always. These methods make foam, which is fine on top of a cappuccino or a hot chocolate and useless for latte art. If you own a machine with a steam wand already, use it instead.

THE JAR. Fill a jar no more than a third full with cold milk, screw the lid on tight and shake hard for about 45 seconds. Take the lid off, then microwave it for 30 to 45 seconds. Shaking cold and heating afterwards is the right order: hot milk in a sealed jar builds pressure, and foam made in already hot milk collapses as fast as you make it. The heat sets what the shaking built.

THE FRENCH PRESS. Pour warm milk into the jug, no more than a third of the way up, and plunge briskly for 30 seconds or so. This gives a finer foam than the jar because the mesh breaks the bubbles down as they pass through it. It is the closest of these to something you would pour.

THE HANDHELD WHISK. The battery-powered wands sold for a few dollars work, and they work better in a tall narrow vessel than in a mug, because the milk needs somewhere to circulate. Heat the milk first, then whisk near the surface.

A balloon whisk and a saucepan does the job too, and takes longer than any of the above. Warm the milk gently, take it off the heat, then whisk hard.

What went wrong

BIG BUBBLES ON TOP. You stayed in the air phase too long, or the tip drifted out of the milk. Cut the hissing to two or three seconds and get the tip properly submerged for the rest.

NO FOAM AT ALL. The tip went in too deep from the start, so you heated the milk and never aerated it. There is no way to add air once the milk is hot, so start again with cold milk.

FOAM SEPARATING FROM THE MILK IN THE JUG. The texturing phase was too short to fold the air through. Swirl the jug hard and tap it once on the counter before you pour, which buys you some of it back, and spend longer on the whirlpool next time.

A COOKED, SCALDED TASTE. You went past about 70 C. Nothing rescues this, and the milk should be poured away rather than served.

PLANT MILK SPLITTING INTO GRAINS. Either it is not a barista edition, or the espresso underneath was very hot and very acidic. Pouring the milk in sooner after the shot helps a little; switching cartons helps more.

Habits worth building

Purge the wand before you start and again the moment you finish, and wipe it with a damp cloth while it is still hot. Milk that dries onto a wand bakes into the holes and eventually blocks them, and by then it needs soaking rather than wiping. This is 5 seconds of work at the time and a job on a Sunday if you skip it.

Fill the jug to the base of the spout, no higher. Milk expands as you aerate it, and a jug filled to the brim gives you no room to work and a mess to clean up.

Steam what you are about to pour. Milk that has been steamed once has already used up much of what makes it foam, so re-steaming leftovers gives a poor result and a slightly cooked taste. Cold leftover steamed milk is fine to drink or cook with, just not to re-foam.

Practise on the cheapest milk you can find, not the good stuff. Getting the air phase timed takes most people a couple of dozen attempts, and on a machine with a manual wand that is a couple of dozen jugs of milk. It is a motor skill rather than a recipe, which is why reading about it only takes you so far.