Yes, alcohol does evaporate when you bake, but rarely all of it. The amount that remains depends on oven temperature, baking time, the surface area of the food, and how the alcohol is bound within the matrix. Most recipes retain somewhere between 5 and 85 percent of the original alcohol, a range so wide that it undermines casual claims that "the alcohol burns off." Understanding why requires looking at the physics of evaporation, the chemistry of alcohol-water mixtures, and the way food structure traps volatile compounds long after the oven door closes.
This matters for more than flavor. People cooking for children, pregnant guests, people in recovery, or anyone avoiding alcohol for religious or medical reasons need a realistic picture. The U.S. Department of Agriculture and major food science bodies have published retention data, and the numbers are more stubborn than most cooks expect.
The evaporation mechanism at the molecular level
Ethanol boils at 78.37 degrees Fahrenheit lower than water's 212 F at sea level. That 35-degree gap is why alcohol leaves first when you cook. At any temperature above roughly 173 F, ethanol molecules at the surface gain enough kinetic energy to break intermolecular hydrogen bonds and escape into the air. Water molecules, held more tightly by stronger hydrogen bonding, need more energy to do the same.
But boiling point is only half the story. Evaporation happens below boiling, driven by vapor pressure. Ethanol has a higher vapor pressure than water at any given temperature, meaning it preferentially enters the gas phase. In an open pan, this drives rapid initial loss. In a sealed cake batter, the ethanol must first diffuse through the batter to reach the surface before it can escape, and that diffusion is slow.
The mixture effect complicates everything. Ethanol and water form a non-ideal solution with strong hydrogen bonding between the two. This interaction raises the effective boiling point of the ethanol fraction and slows its release compared with pure ethanol. A wine at 12 percent alcohol does not behave like 12 percent pure ethanol; the water matrix holds onto the alcohol longer than simple math predicts.
A second mechanism matters inside baked goods. Alcohol can become physically encapsulated in starch matrices during gelatinization, or bound to proteins and fats. Ethanol is both water-soluble and fat-soluble, so it partitions into whatever phase is available. In a butter-rich batter, a meaningful fraction dissolves into the fat phase, where it is shielded from airflow and evaporates far more slowly.
How much alcohol actually survives: quantified context
The most-cited retention figures come from USDA-funded research by Augustin and colleagues, published in the 1980s and still used as the reference standard. These numbers describe the percentage of alcohol remaining after a dish is cooked and served immediately, and they are the source of most retention guidance in food science.
| Preparation method | Alcohol retained | Notes |
|---|---|---|
| Alcohol stirred into boiling liquid, removed from heat | 85% | Brief contact, no sustained simmer |
| Flambé with no residual liquid | 75% | Flame does not consume as much as assumed |
| Baked or simmered 15 minutes | 40% | Common marinade and pan-sauce scenario |
| Simmered 30 minutes | 35% | Sauce reductions |
| Baked 1 hour | 25% | Long braise or casserole |
| Baked 2 hours | 10% | Extended oven time |
| Baked 2.5 hours or more | 5% | Near-complete but never zero |
The table shows a clear trend: time dominates over temperature. Doubling cooking time reduces retention roughly by half in the early stages, then flattens as the residual alcohol becomes trapped. The 5 percent floor persists because some ethanol remains bound in the food matrix even after hours of heat.
These are averages. Real retention can swing widely based on pan geometry, whether the dish is covered, the ratio of surface area to volume, and the food's moisture content. A thin layer of sauce in a wide skillet loses alcohol far faster than the same volume in a deep, lidded pot. A dense cheesecake retains more than a porous sponge.
For practical perspective, a dish retaining 25 percent of the alcohol from a cup of wine at 12 percent alcohol delivers roughly 0.5 fluid ounces of pure ethanol per serving if the recipe serves four. That is comparable to a small fraction of a standard drink, but for someone strictly avoiding alcohol it is not zero.
Edge cases where the numbers break down
Several situations push retention far from the textbook averages, and these are where thin explainers fail.
Alcohol injected or soaked into a dense food retains much more than surface-applied alcohol. A rum-soaked cake, a beer-braised roast, or a wine-marinated steak carries alcohol into the interior, where temperatures may not exceed 160 F for hours. Interior temperatures in a large roast can lag the oven by 100 F or more during early cooking. The surface evaporates fast; the center does not.
Covered cooking traps vapor. A lidded Dutch oven or a foil-covered casserole allows ethanol to condense on the lid and drip back into the food, effectively recycling it. Retention can be double the open-pan figure for the same time and temperature.
Added alcohol late in cooking barely evaporates. A splash of wine added at the end for deglazing or finishing retains a large fraction of its ethanol, often 50 to 70 percent, because there is not enough time or surface exposure for meaningful loss.
Alcohol in no-bake or refrigerated recipes does not evaporate at all. A rum ball, a tiramisu, or a no-bake cheesecake with spirits retains essentially 100 percent of the added alcohol. This is the most common source of accidental exposure and the one most often overlooked.
The flambé assumption is the most persistent error. Lighting alcohol on fire does not incinerate it; the flame burns a fraction of the ethanol vapor above the surface while much of the liquid remains. Retention after flambé is typically three-quarters of the starting alcohol, not the "all burned off" claim repeated in cooking shows.
Warning: no-bake desserts with spirits retain essentially all added alcohol. Warning: flambé removes only about 25 percent, not all, of the alcohol. Warning: alcohol added at the end of cooking retains 50 to 70 percent. Warning: covered cooking traps ethanol vapor and recycles it into the food. Warning: interior alcohol in large roasts or soaked cakes survives long baking.
Misconceptions that persist in cooking media
The claim that alcohol "cooks off" in a few minutes is the most widespread myth. The Augustin data show that a 15-minute simmer leaves 40 percent, and only after two and a half hours does retention approach 5 percent. Cooking shows and recipe blogs routinely understate this.
A second misconception treats alcohol as a single substance. Beer, wine, and spirits differ in alcohol concentration and in the presence of sugars and other dissolved compounds that affect evaporation. Wine at 12 percent and vodka at 40 percent do not lose alcohol at the same rate per unit volume, though both follow the same basic physics. The dissolved solids in wine also raise the mixture's boiling point slightly, slowing water loss and indirectly slowing alcohol loss.
A third error confuses evaporation with elimination. Ethanol that leaves the food enters the kitchen air. In a small, poorly ventilated kitchen, cooking with a large volume of wine or spirits can raise indoor ethanol concentrations noticeably. The amounts are generally small and short-lived, but the point is that the alcohol is not destroyed; it is relocated.
A counterintuitive detail: ethanol is more volatile than water, yet in a simmering stew the alcohol does not simply disappear first. Because ethanol and water hydrogen-bond, the ethanol rides out with water vapor in a roughly proportional stream. The mixture loses both together, which is why long cooking is required to push retention below 10 percent. The stubborn residue is not a failure of physics but a consequence of it.
Professional bakers know a related caveat. In yeast-leavened bread, fermentation produces ethanol, and some of it survives baking. A loaf can retain a small but measurable fraction of fermentation ethanol, typically well under 0.5 percent by weight, because the crumb structure traps it. This is one reason commercial breads are not entirely alcohol-free, though the levels are far below anything intoxicating.
Practical actions for controlling alcohol retention
The following steps translate the physics into kitchen decisions with concrete thresholds.
For a dish you want to taste of wine without carrying much alcohol, add the wine early and simmer uncovered for at least 30 minutes. Retention drops to roughly 35 percent at 30 minutes and about 25 percent at one hour.
For a sauce that needs a wine note, add half the wine at the start and reduce; add the rest off heat if you want aroma. A late addition preserves the volatile aromatics that long cooking destroys.
For people avoiding alcohol entirely, do not rely on cooking to remove it. Substitute an alcohol-free wine or a stock-acid combination (broth plus a splash of vinegar or lemon) rather than assume 5 percent residual is acceptable. For strict avoidance, 5 percent is not zero.
For large roasts and soaked cakes, expect interior alcohol to persist. If the dish must be alcohol-free, do not use a marinade or soak with spirits; the interior will not reach evaporation conditions.
Ventilate the kitchen when cooking with substantial volumes of wine or spirits. Run the range hood and open a window during reductions. The ethanol leaving the pan is real, and in a small kitchen, a quart of wine reduced over 20 minutes releases a measurable amount into the air.
Check the recipe's timing against the retention table. A "5-minute wine sauce" retains most of its alcohol. A "2-hour braise" retains little. The recipe's stated time is the best predictor of what remains.
What the data can and cannot tell you
The Augustin retention figures are averages from controlled experiments, not guarantees for any specific recipe. They assume a particular pan geometry, a particular heating profile, and immediate serving. Variation is real and can be large.
Retention also depends on how the dish is served. A stew served with the cooking liquid retains more alcohol than the same stew with the liquid drained. A cake brushed with syrup retains more than one left plain. These serving choices are rarely specified in recipe instructions, so the cook must reason from mechanism.
The 5 percent floor is an average, not a limit. Some dishes, particularly those with high fat content or dense starch matrices, retain more even after long cooking. Others, thin and uncovered, drop below 5 percent. There is no reliable way for a home cook to measure residual alcohol with kitchen equipment; the data give a range, not a number.
One thing the data do establish firmly: alcohol does not fully evaporate in any normal cooking scenario. Claims that it does are not supported by the food science literature. For most cooking, the residual amount is small and generally considered safe for adults. For anyone who must avoid alcohol for medical, religious, or recovery reasons, the only reliable strategy is substitution, not extended cooking.
The question "does alcohol really evaporate when you bake?" has a precise answer: it evaporates partially, predictably enough to model, but never completely. The oven is not an incinerator, and the food matrix is not a sieve. What remains is a small, measurable, and often manageable fraction that depends on time, surface area, structure, and whether the pan is covered. Knowing those variables lets you cook with wine and spirits deliberately rather than on myth.