Browning and the Maillard Reaction

What the Maillard reaction actually is

Amino acids and sugars, under heat, recombine into hundreds of aroma compounds. Roast meat, bread crust, coffee, soy sauce — all of it comes from this reaction. It is why boiled meat and seared meat are different foods.

There is one condition that matters: the surface has to get well above the boiling point of water. While there is water on the surface, the temperature stalls at 100°C and essentially no browning happens.

"Essentially" is doing real work in that sentence. Maillard runs below 100°C too — just very slowly — so time can stand in for heat. Pumpernickel bakes a wet dough in a covered tin, never above 100°C, for sixteen hours, and comes out black. Black garlic and dulce de leche are the same trade.

So "well above" is the condition for doing it in minutes. In a pan you buy browning with temperature; given an oven and a night, you can buy it with time instead.

There is a second condition: pH

Maillard runs considerably faster in alkaline conditions. So colour can be bought with pH instead of only with temperature and time.

Bagels put soda in the boiling water, because at a high pH the crust browns faster and glossier, and Pretzels push the same principle as far as lye. Both of those are alkali in a dough or a bath.

Buffalo Wings paint it straight onto the skin. Baking powder rubbed on raises the skin's pH, which speeds the browning and at the same time loosens the skin's protein structure so water escapes instead of steaming underneath. That is how a 220°C oven produces a crust that would otherwise need a fryer.

And a third: what the surface is made of before it meets heat

Temperature, time and pH are all ways of pushing the same reaction harder. The composition of what is reacting is the other half, and the cheapest way to change it is fermentation.

A miso glaze scorches in thirty seconds under a grill that would leave a raw surface pale, because a year of fermentation has already cut the protein into free amino acids and the reaction has nothing left to wait for — fermented pastes treats that as the central hazard of cooking with them. The sugar side is the same lever from the other end: what sugar does notes that a little sugar rubbed on meat browns it faster because it gives Maillard more to work with, which is why a yoghurt-marinated kebab chars while a bare skewer is still pale.

So before reaching for more heat, ask whether the surface already has both halves of the reaction in a usable form.

When one half is missing, no lever reaches it

Everything above adds to the surface, so composition reads as a way of having more of the reaction to work with. The same axis runs the other way, and then it is not a lever at all: if one of the two halves is not on the surface, temperature, time and pH have nothing to push. The colour has a ceiling, and more heat does not raise it.

• The protein side missing. Karaage is dredged in potato starch and nothing else, and its page draws the conclusion citing this guide — "potato starch supplies none, so karaage stops at light gold. Deep brown here is the soy in the marinade caramelising, not the coating browning". The second fry still does the rest of its job, driving the last water off the surface and setting a thin glassy shell; what it does not do is this reaction. So colour goes out of service as a reading, and that page's instruction is "Go by temperature and time instead." Tonkatsu sits at the other end of Kushikatsu's lineage — "Tempura is batter with no crumb at all and karaage is starch only" — where flour, egg and breadcrumb put protein at every stage, and it is fried "until golden and crisp" because there the colour can be read.
• The heat side missing, usually on purpose. Tortilla de patatas holds its oil at 85-95°C for the whole twenty minutes, citing Carnitas: "the surface never dries and never browns … the colour never arrives." No colour is the design there, because "A browned edge is a dried, firmer surface" and this slice has to stay soft to its face so the egg meets it rather than running around it. Tempura holds the same condition for most of a fry rather than all of it — at 175°C "Most of that time goes on driving the water out of the batter rather than on cooking the ingredient", and while that water is there the surface is pinned at 100°C, which its page says citing this page.

None of that is a reaction running ahead of you, which is the other reason colour fails as a reading. What sugar does judges a caramel by temperature because "By the time you see the colour you want, you are late", and cooking oils gives meunière a thirty-second window where "colour arrives too late to judge by, the cue is the smell". There the colour is real and behind you; here it is not coming at all. Thickening a sauce has a case of its own again — a starch film that is not allowed to colour, kept wet and finished in liquid, so it never gets the chance rather than lacking the means.

So the first thing is dryness

• Pat the surface completely dry with paper towel.
• Do not go straight from a wet package in the fridge into the pan.
• Marinated meat has to be shaken off — soy-based marinades are especially wet. A thick coating is the opposite case and is meant to stay on (what marinades do).
• Frozen meat needs to be fully thawed, then dried.

Those four lines read as things you do before the pan, but the same fact is the reason for several steps that look nothing like patting something dry. "While water is present it stalls at 100°C" is a constraint on the order of work, not a prep step.

• If anything wet is going on top, the browning has to happen first. Croque Monsieur toasts the bread in a pan and only then takes its béchamel and its second heat, because the two browning events cannot be combined. Gyoza is the case that has to survive the opposite order: the base is fried to a crust, then water pins it back at 100°C and softens that crust, so it is fried a second time once the water has gone.
• The water can be hidden inside the food. That is why the potatoes in Patatas Bravas are rinsed. Surface starch gelatinises in the heat into a gummy film, and that film holds water against the potato; rinsing the starch off is not tidiness, it is removing the film.

The second thing is not crowding the pan

This is the most common failure of all. Fill the pan and the moisture coming off each piece cannot escape; it drops the pan temperature, and you get grey, steamed meat instead of a crust.

You should see gaps between the pieces. Cook in two or three batches. It feels like twice the work, but meat that has steamed will never brown no matter how long you leave it. That is why the recipes for Pique Macho and Beef Stroganoff both insist on two batches.

The third thing is leaving it alone

Constant flipping cools the contact surface, so there is never enough time for a crust. Sear one side properly and turn it when it releases from the pan by itself. If it sticks, it is not ready.

That reads as a rule about meat, and it has nothing to do with protein. Polenta is a set starch with no fibres and no connective tissue, and touching a face before it has set on the metal tears that face away — four or five undisturbed minutes is the whole technique. The condition is not the meat, it is whether the surface has finished giving up water and set.

The brown film in the pan is an ingredient

After the meat comes out, that brown layer stuck to the base — the fond — is concentrated Maillard product. Washing it away throws out half the flavour.

And if the fond is the point of searing, then "sear it to lock in the juices" is not true. A browned surface is porous rather than a seal, and seared meat loses slightly more water than unseared. Lomo Saltado faces that head-on: its beef is juicy because the total time on the heat is about three minutes, not because a crust is holding anything in.

Deglazing: with the pan still hot, add liquid and scrape the base with a wooden spoon.

• Wine or beer: acid and alcohol lift it well. Beef Stroganoff, Pique Macho.
• Stock or water: neutral and reliable.
• Vinegar or lemon: powerful in small amounts.

Reduce that liquid by half and it is already a sauce.

Caramelisation is a different reaction

Both go brown, but the mechanism differs. Caramelisation breaks down sugar alone (sugar, onion, carrot); Maillard needs sugar and protein together. The 40 minutes of onions in French Onion Soup is caramelisation; the browned cheese on top is Maillard.

There is a third browning, and it is the one that needs no heat. It is what turns a cut apple or potato brown within minutes: enzymatic browning, polyphenol oxidase meeting oxygen at the cut surface. It has nothing to do with a flame and shows in under a minute.

Some steps only make sense once you know it. Soaking cut burdock in Kinpira Gobo is not about rinsing starch off the way it is with potato — it is about excluding oxygen — which is also why more than five minutes is a loss, since the aroma leaves for the water too. Gotgam is the same reaction run to completion over weeks: the astringent tannins link into molecules too large to grip the tongue, so the astringency alone disappears and the sweetness reads.

Where you can break that reaction is decided by what else the surface is for

The paragraph above gives one answer — keep the oxygen away — and it is the one that comes to mind first. But the reaction needs the enzyme, something for it to act on, and oxygen, all three at the cut surface at once, and removing any of them stops it. Which one you can remove is not a matter of preference.

• Take the oxygen away. Water is the usual cover, and Yukhoe uses it on the pear it slices alongside: "soak it in cold water for 5 minutes, then drain. That stops it browning and takes the harshness out." A fat does the same job and wets nothing, which is why Pesto pasta keeps leftovers "with a thin film of olive oil on top to slow browning". The beef on that same yukhoe plate is a different chemistry — myoglobin rather than an enzyme — and answers to the identical move: "a film of oil over the surface reduces contact with oxygen, so the myoglobin browns more slowly."
• Take the enzyme away. Acid slows it — kinpira gobo's optional vinegar in the soaking water "stops the browning more decisively and takes some of the earthy aroma with it", and the lime in Guacamole is doing that as well as brightening it. Heat destroys it outright, which is why Lahpet thoke begins by steaming: "left alone after picking, tea leaves oxidise by their own enzymes and become black tea. Steaming kills those enzymes so the green colour and the grassy note survive." Green tea and black tea are one leaf and that one step.
• Take neither, and cut later instead. Gamja-jeon writes the bind down: "water would keep the oxygen out, but it would also wash the starch away, so here the answer is not to grate it early rather than to soak it." That starch is the pancake's only adhesive, so the cover the oxygen route offers costs more than the browning does.

Read that way, the burdock's five-minute ceiling above is the general case rather than a quirk: a cover that works by being a solvent takes things away as well as keeping oxygen out (soaking), so where the surface has a job of its own, starting the clock later often costs less than slowing it down.

Browning used repeatedly, not once

In everything above, the brown layer is something you protect — leave it until it releases, then scrape the fond up as a sauce. Some dishes deliberately break it and rebuild it.

• Cassoulet pushes the crust that forms on top back under with the back of a spoon. In Castelnaudary that is done up to seven times, and only the last one is kept. Every broken crust returns its browned surface and its fat to the pot, so the finished dish tastes of its own crust.
• Baumkuchen browns one thin layer of batter under direct heat, pours the next over it, and repeats about twelve times. Other cakes bake as a mass, so only their crust browns; here you eat twelve browned surfaces at once. The rings and the flavour are two faces of one fact.

And browning can make sweetness with no sugar involved. Brunost boils whey down for hours with not a grain of sugar added: its lactose and whey protein brown against each other, which is why it ends up deep brown and tastes closer to fudge and nuts than to milk.

Frying obeys the same rule

This is the reason Tangsuyuk and Karaage are fried twice. The first fry cooks the inside and drives moisture out of the surface; only in the second, hotter fry does the surface get well above 100°C and the crust set golden and crisp. Whether that colour is this reaction depends on what the coating is made of, though: both of these coatings are potato starch, which is the case the section above is about. Fry once and you either burn the coating while the inside cooks, or pull it out pale.