Fat Washing

Fat washing is a finishing step, not an aging step. It goes last — after the oak comes out, after any resting, right before you bottle. Oak and fat are competing for the same part of your palate and the same weeks of your patience, and the fat will happily strip some of what the wood just gave you (more on that below). Get the spirit where you want it on wood first, then decide whether it wants a wash.

Stir melted fat into whiskey and leave the two sitting together. They never blend — but the flavor in them moves. Over a few hours it drifts across the boundary between the fat and the whiskey until it settles. Freeze the fat hard, lift the solid cap off, and whatever crossed over stays in the glass.

The part almost nobody mentions is that the traffic runs both ways. While the fat is handing over its bacon or brown-butter flavor, it is also quietly absorbing some of the whiskey’s own — and everything it absorbs goes in the bin with the cap.

So the outcome is a trade, not an addition. Your whiskey gains the rich, savory, cooked character of the fat. It also gives up a little of what it came with: some of its oak and fruit notes, and some of its heat and grip. Most people taste the result as rounder and softer — and the likely reason isn’t that something was added to smooth it, but that something was taken away.

The alcohol is the bridge. Water and fat want nothing to do with each other, so on their own almost no flavor would cross. Alcohol gets along with both, which is what lets the flavor travel — and you need a good deal of it. Below roughly 17% alcohol, barely anything moves. That’s why you can fat wash whiskey, rum or gin but not wine or beer, and why a cask-strength bottle pulls harder than a 40% one.

The grease can’t follow the flavor. Flavor molecules are only mildly fat-preferring; a molecule of fat itself is overwhelmingly so, and it simply won’t dissolve into a cold spirit. Even pure alcohol barely dissolves oil at room temperature — oil doesn’t properly blend in until around 70 °C. That gap is the whole trick: the taste crosses over, the oiliness can’t. It’s also a good reason not to warm the jar to hurry things along, since heat is one of the few things that does pull fat into alcohol, leaving you more to freeze back out.

Which flavors cross over and which stay put is predictable, and there’s a number for it: log P, a measure of how strongly a compound prefers fat over water. Sort the compounds in play by it and the whole technique becomes legible. Everything near the top of this list stays in your glass no matter what you do. Everything near the bottom rides out on the fat cap. The dividing line isn’t arbitrary either — at spirit strength, a fat phase measurably holds on to compounds above about 2.5 and leaves the ones below it alone. What’s still predicted is the last column: nobody has measured these particular compounds in an actual fat-washed spirit.

log P values are PubChem’s computed XLogP3 estimates — trust the ranking, treat the decimals as approximate.

What arrives in the glass is not “bacon” or “butter” in the abstract — it’s a specific short list of molecules, and both have been characterised properly.

• Bacon fat brings hexanal (green, grassy), 1-octen-3-ol (earthy, mushroomy) and 2,4-decadienal (the deep-fried, oily note) as its loudest voices, with pyrazines carrying the roasted, meaty character underneath. The savory side is Maillard chemistry; the fatty side is fatty-acid breakdown.
• Butter is simpler than it smells: of 230+ identified volatiles, only a handful matter. Diacetyl does the buttery lifting, butanoic acid the cheesy edge, and δ-decalactone a soft stone-fruit note. Browning the butter first adds a whole second layer of Maillard compounds — which is why brown butter outperforms plain butter at the same weight.
• A counterintuitive one: in non-smoked bacon, over 88% of the aroma compounds live in the lean, not the fat — 127 flavor precursors in the meat versus 26 in the fat. Rendered fat is mostly a solvent carrying flavor off the meat, not the source of it. Practical upshot: render your fat from properly browned bacon and don’t strain the fond out too fastidiously. How brown the bacon got matters more than how much fat you collected.

Every article on fat washing promises a rounder, silkier texture from lipid left behind in the spirit. The measurements don’t just fail to support that — they argue against it. Nobody has ever quantified residual lipid in a fat-washed spirit, so any figure you see quoted is invented. But we can bracket it, and the brackets are damning:

• Lubrication needs about 1% fat. In milk tested from 0.06% to 8% fat, friction dropped sharply — and creaminess was actually perceived — only above roughly 1% fat. Below that, added lipid does very little.
• You’d have to taste it, and you can’t. Free fatty acids become detectable around 220 mg/L. The long-chain esters in a real whiskey sit at 0.1–1.5 mg/L — about two orders of magnitude below threshold.
• Coating and viscosity are both out of reach. It takes a ~25 µm change in the fat film on your tongue to shift perceived mouthfeel, and roughly a 26% change in viscosity to notice one at all. A spirit is about as viscous as water.

So the sensation people report is probably real — the explanation isn’t. And there’s a much better one available, which is the next section.

The interesting possibility is that fat washing doesn’t soften a spirit by adding anything. It softens it by taking something out. Three separate published findings line up into a chain:

1. A lipid phase selectively grabs the hydrophobic congeners. Add a fatty ester to aqueous ethanol at spirit strength and it suppresses the headspace concentration specifically of compounds above about log P 2.5, leaving the water-soluble ones alone. That’s the fat acting as a selective sponge — measured, in an ethanol matrix.
2. Those trace congeners are what make a spirit feel hot. In a neat piece of study design, rum was diluted two ways: with water, and with 40% ethanol so the proof stayed put while the congeners halved. At identical ABV, halving the congeners dropped perceived warming from 9.2 to 3.8 and astringency from 9.8 to 5.6. Watering it down didn’t significantly reduce warming. The heat is in the trace compounds, not the alcohol.
3. And the specific burn-drivers are fat-loving. The compounds that potentiate ethanol burn are lipophilic carbonyls — octanal, nonanal, benzaldehyde, 2-heptanone — with octanal registering at 0.05 ppm. Exactly the molecules a fat phase should pull out.

Every link in that chain is published. The only step nobody has taken is the last one: actually measuring it on a fat-washed spirit. So this stays a well-supported hypothesis rather than a finding — but it explains the reported smoothness far better than a film of grease does, and it reframes the technique as something closer to a filter than an infusion.

This is the half of the technique the internet skips. Practically every source claims fat washing only adds — and none of them measured it either. Equilibrium doesn’t work that way: the fat is an extractant, and the moment it goes into the jar it starts pulling your whiskey’s own lipophilic congeners out of solution. When you lift off the frozen cap, they go in the bin with it.

There is one very well documented precedent for exactly this. Chill filtration — the industrial process that removes long-chain fatty-acid ethyl esters from whisky to stop it hazing — is essentially the same physics running on the same molecules. And note what you do in step three of a fat wash: you put a whiskey in a freezer. Spirits held at −18 °C measurably cloud up at 40% ABV while staying clear at 70%, because those esters come out of solution as the temperature and the alcohol fraction drop. Then you filter. Cold filtration removes over 90% of a marker ester where the same filter at room temperature removes about a third — so filtering cold works precisely because it is chill filtration. You get the clarity and you pay for it in the compounds our Body & Mouthfeel page tells you to protect. There’s no version of this where you keep both.

Ranked by how hard the fat should pull, based on log P:

• Long-chain fatty-acid ethyl esters (ethyl laurate 5.6, ethyl palmitate 7.8) — the heavy, oily esters behind a spirit’s body. Highest log P in the glass; first to go.
• Oak lactones (2.5) and eugenol (2.0) — your coconut, fresh-wood and clove notes. Squarely in the range where meaningful transfer is expected. The single best reason to fat wash after you’ve dialled in your oak, and to taste before and after.
• Mid-chain fruity esters (ethyl octanoate 3.5, ethyl decanoate 4.6) — some loss expected.
• Vanillin (1.2) and guaiacol (1.3) — too water-soluble to care. Your vanilla and your smoke should come through essentially untouched.

Four steps and one overnight wait. You need a jar with a lid, a freezer, and a coffee filter.

1. Render and cool the fat. Cook the bacon well browned, or brown the butter, then pour off the liquid fat and let it cool to just-warm. Straight from the pan is too hot — you’ll flash off both the ethanol and the most delicate aromatics, and hot fat plus a high-proof spirit is a genuine fire risk. Use unsalted butter and low-sodium bacon; salt crosses into the spirit far more readily than flavor does.
2. Combine and infuse. Add the fat to the spirit in a sealed jar and shake. Leave it at room temperature for 4 hours to overnight, shaking whenever you walk past. The shaking isn’t superstition: extraction is limited by how much surface area the two phases share, so breaking the fat into small droplets genuinely speeds it up. Let it stand and the droplets merge and the process slows.
3. Freeze it solid. Into the freezer for at least 2 hours; overnight is better. A 40% ABV spirit freezes around −23 °C and a home freezer runs about −18 °C, so the spirit stays liquid while the fat sets hard on top. That gap is the whole trick — and it’s why a low-ABV base (under ~25%) will slush on you.
4. Lift the cap and filter cold. Break off the solid fat and discard it, then pour the spirit through cheesecloth, and then through a coffee filter. Do this while everything is still ice cold, working fast, and put the filter and receiving bottle in the freezer beforehand if you can. This is the step most people rush, and it’s the one with the hardest numbers behind it: the same filter that catches about a third of a marker ester at room temperature catches over 90% of it under freezing conditions. Warm up first and the sub-visible droplets that slipped through will coalesce into a haze later. A second freeze-and-filter is normal, and with liquid oils, a third.

There is no correct ratio, and anyone who gives you one number is overselling. Published practitioner ratios span from about 1 part fat in 17 to 1 in 2 — a nine-fold range — because the right amount depends entirely on how loud the fat is. The rule that actually holds: the more assertive the fat, the less of it you need. Start low. You can always run a second wash; you can’t un-wash a bottle.

• Why proof isn’t the protection you think. Ethanol only becomes a reliable preservative in its own right somewhere around 18–21%, and a 40% spirit does clear that bar as poured. But that’s not where the risk lives. It’s in how you handled the fat before it went in, in any dilution afterwards, in a low-ABV base — and in rancidity, which alcohol does nothing about at all.
• Rancidity is the realistic failure. Residual lipid oxidises exactly as it would anywhere else — a free-radical chain reaction driven by oxygen, light, heat and metal, faster the more polyunsaturated the fat. Ethanol is not an antioxidant. Store cold, dark, and in a bottle with little headspace.
• Meat fats are the stricter case. Cook the meat through before rendering, keep the whole process cold, and don’t store the finished bottle anywhere you wouldn’t store raw meat. Plan on days for animal fats — a week or two at the outside — and up to about a month for plant oils. Those numbers are practitioner consensus and they disagree with each other; nobody has actually measured this, so err short.
• Never wash with a home-made fresh garlic or fresh herb oil. Oil plus fresh, low-acid aromatics is the classic anaerobic setup for Clostridium botulinum — real enough that the FDA has required commercial garlic-in-oil to be acidified since 1991. Making such an oil safely (thoroughly dried aromatics, or acidification, or same-day refrigerated use) is a prerequisite before it goes anywhere near your spirit, not something the alcohol fixes afterwards.
• Does it lower the proof? Slightly, probably, and nobody has measured it. The discarded fat cap traps some liquid, and ethanol partitions into fat rather more willingly than water does — so the cap should carry off proportionally more alcohol than water. At normal ratios you’re talking about a few percent of volume. Don’t trust anyone who states flatly that the ABV is unchanged.

It isn’t a bartender’s invention so much as a borrowed one. Perfumers in Grasse were doing the second half of it in the 1700s: enfleurage laid flowers on odourless fat for weeks until the fat was saturated with their scent, then washed the fat with alcohol to pull the fragrance back out. Same two phases, same partition, same direction of travel — three centuries earlier.

The modern version came out of wd~50 in New York in the mid-2000s, credited to Tony Palomino, who spread peanut butter on a sheet pan and poured a thin layer of spirit over it overnight. Eben Freeman carried it forward and taught it to Don Lee, who in 2007 rendered Benton’s hickory-smoked bacon fat into Four Roses and put the Benton’s Old Fashioned on the menu at PDT. His own framing is still the cleanest definition anyone has managed: “All fat washes are infusions. Not all infusions are fat-washing.”