Why does a drop of diluted whisky dry into a web? Marangoni flow, monolayer collapse, and the pattern that only appears near 20% ABV

Key takeaways

• A “whiskey web” is a microscopic net of strands left when a 1.0 µL drop of diluted American whiskey evaporates on glass. It forms from nanoscale agglomerates that stratify at the liquid–air interface and then collapse as the drop shrinks (Carrithers et al., ACS Nano, 2020).
• The pattern is a dilution effect. Above about 35% ABV the drop dries to a uniform film; near 10% ABV it dries to a coffee ring; the web appears in between, at roughly 20–25% ABV (Science News).
• It is also a whiskey-type effect. Of 66 off-the-shelf American whiskeys, “all but a 42-year-old sample formed webs at 25% ABV”, while Scotch and Canadian whiskies do not web under the same conditions (Physics Today, 2021).
• The webs are brand-specific enough to be forensic. Image processing matched unknown samples to the right product around 90% of the time across 15 droplets per whiskey (Physics Today, 2021).
• Nothing in the web comes from the water — it is built from the spirit’s water-insoluble congeners. Any non-volatile solid in the diluent would be deposited alongside them, which is why laboratory dilution protocols specify distilled or ultrapure water (Baxter, Réveillé & Conner, J. Inst. Brew., 2024).

A drop of diluted American whiskey dries into a web because dilution pushes the spirit’s water-insoluble compounds out of solution and into a skin at the drop’s surface, and the evaporating ethanol then buckles that skin until it folds and collapses onto the glass. The effect is strongly strength-dependent: it appears at roughly 20–25% ABV and disappears at cask strength or at high dilution. Everything visible in the pattern came out of the whiskey, not out of the water — which makes the phenomenon a neat demonstration of what reduction water is supposed to be, and what happens when it is not.

What is a “whiskey web”?

It is a microstructure, not something you can see in a glass. Researchers at the University of Louisville deposit a 1.0 µL droplet of diluted whiskey on a cleaned glass coverslip, let it evaporate at ambient conditions, and photograph the residue under a microscope. What they found in American whiskeys was a lattice of fine strands — reported as a previously undocumented phenomenon and named the whiskey web (Carrithers et al., ACS Nano, 2020); the first published images appeared as a Gallery of Fluid Motion entry, Whiskey webs: Microscale “fingerprints” of bourbon whiskey (Williams, Brown & Carrithers, Phys. Rev. Fluids 4, 100511, 2019).

The formation sequence set out in the accompanying thesis work runs in five stages: “(1) dilution of the whiskey to form nanoaggregates, (2) formation of a monolayer at the air-liquid interface, (3) chaotic ethanol evaporation caused monolayer collapse (via dynamic pressure), (4) bulk fluid evaporation caused monolayer collapse (via reduction of surface area), where the web-like structures reside on the surface, and finally, (5) web-like structures deposit on the substrate” (Brown, University of Louisville, 2019). Under the electron microscope the folds resembled the “twisted ribbon fold” already described in the monolayer literature.

Two numbers make the mechanism concrete. The strands are the wreckage of a surface film, and that film is destroyed largely by geometry: as the drop evaporates, its surface area falls by roughly 30%, and the monolayer riding on that surface has nowhere to go but into folds.

Why does the pattern only appear when the whisky is diluted?

Because dilution is what creates the particles in the first place. Whisky carries long-chain fatty acids and their ethyl esters in solution at high strength; add water and they stop being soluble. The same chemistry that produces visible louching in a glass — “polarisation causes the molecules to clump together to form structures known as micelles”, which precipitate as “the proportion of water in whisky is increased (the critical point seems to be about 46% alcohol by volume)” (Annandale Distillery) — produces the colloidal clusters that a whiskey web is built from. We set out the visible version of this in why whisky goes cloudy when you add water.

The strength window is narrow and well documented. Bourbons at 35% ABV and above left “uniform residue films”; at about 10% ABV they left “markings similar to coffee rings”; and almost every American whiskey diluted to around 20% alcohol left the weblike microstructure (Science News). The published ACS Nano work puts the effect most strongly in the 20–25% ABV band.

That band should look familiar. It is the same dilution the Scotch industry uses for quality nosing, for reasons that have nothing to do with microscopy and everything to do with the nose — as covered in why tasting panels dilute whisky to 20% ABV. Two different disciplines arrived at the same strength because that is where the dissolved, invisible part of a whisky starts behaving like suspended matter.

Why does coffee leave a ring and whisky does not?

The default behaviour of a drying drop is the coffee ring: liquid evaporating fastest at the pinned edge draws an outward capillary flow that carries suspended particles to the rim, leaving a dark annulus and a clear centre (Deegan et al., Nature 389, 827–829, 1997).

Whisky resists that, and the reason was worked out at Princeton before anyone had seen a web. Tracking the flow inside evaporating whisky droplets, the team found two ingredients doing the work: fat-like molecules that lower surface tension and generate an inward Marangoni flow opposing the outward capillary flow, and plant-derived polymers that stick particles down where they land (Kim et al., Phys. Rev. Lett. 116, 124501, 2016; Phys.org). The result is an unusually even deposit — interesting enough that the authors proposed it as a route to uniform industrial coatings (APS Physics).

The web is what happens when you take that same Marangoni machinery and add enough water to make the cargo insoluble. “Marangoni flows produced by the differential evaporation of ethanol in the droplet drive colloidal clusters—aggregates of insoluble congeners—to the liquid–air interface”, where they assemble into a monolayer that then breaks into chains (Physics Today, 2021).

Do all whiskies web?

No — and the exceptions are as interesting as the rule.

• American whiskey webs almost universally. In the largest published survey, “66 off-the-shelf American whiskeys, 56 of which were bourbon whiskeys. All but a 42-year-old sample formed webs at 25% ABV” (Physics Today, 2021). Maker’s Mark, Pappy Van Winkle’s and Jack Daniel’s are named among them (Science News).
• Scotch and Canadian whisky do not, at the dilutions and droplet sizes used (Science News).
• Change the conditions and the line moves. A Canadian whisky and an Irish one “formed structures in a 2 µl droplet at 40% ABV” (Physics Today, 2021) — so this is a matter of which compounds are present at what concentration, not a hard national boundary.

The prime suspect for the American bias is the barrel. American straight whiskey is matured in new charred oak, which loads the spirit with extractives that a refill cask will not; consistent with that, more lines form when a whiskey is spiked with lignin, the oak-derived polymer associated with barrel maturation (Physics Today, 2021). Heavy filtration cuts the other way, by removing the very colloids the web is built from.

Can a web identify a whisky?

Reasonably well. Because each product carries its own mix of congeners, each produces its own pattern — as one of the researchers put it, “they’re all going to bend and fold in different ways” (Science News). Using digital image processing across 15 droplet samples per whiskey, unknown samples were correctly matched around 90% of the time (Physics Today, 2021), and the ACS Nano authors put the technique forward as a candidate for product authentication and counterfeit detection.

Ninety per cent is a research result, not a court-admissible standard, and it needs a reference library of genuine samples to match against. For provenance questions the analytical mainstream still runs on isotope ratios and congener profiling — including the isotopic signature the proofing water itself leaves behind, which we cover in the isotopic fingerprint of blending water.

What does any of this have to do with the water?

Everything, by omission. The web is described in the literature as a structure of the whiskey’s water-insoluble chemical constituents. Water is the trigger and the vehicle; it is not an ingredient in the deposit, because pure water evaporates and leaves nothing.

That is a property of pure water only. Dissolved solids do not evaporate — whatever minerals a diluent carries are still on the coverslip when the ethanol and the water have gone, adding crystalline deposit that has nothing to do with the spirit. This is why published sensory and analytical protocols do not dilute samples with tap water: they specify distilled water (Baxter, Réveillé & Conner, J. Inst. Brew., 2024) or ultrapure 18.2 MΩ·cm water (Ashmore et al., Foods, 2023). A diluent that contributes solids contaminates the measurement.

Scale the coverslip up to a proofing tank and the logic is unchanged, minus the evaporation. Reduction water is a third or more of a finished bottle, nothing in it leaves again, and the tolerances are small: “as little as two ppm of calcium in the final product can cause precipitation”, which is why reduction water “must be purified, either by reverse osmosis (RO) or deionization” (Artisan Spirit). The microscope simply makes the point photogenic — see what is blending water for the production version, and tap vs filtered vs mineral vs demineralised for what each alternative leaves behind.

Can you see this at home?

Not the web itself. A 1.0 µL droplet is about a fiftieth of a drop from a pipette, and the strands are microscopic; the published images come from optical and scanning electron microscopy, not from a glass left on a table.

What you can observe is the chemistry one step earlier. Take a cask-strength or non-chill-filtered whisky, add water a few drops at a time, and watch the spirit turn faintly hazy as it passes roughly 46% ABV — those are the micelles (Annandale Distillery) that a whiskey web is later assembled from. Use water that is genuinely tasteless when you do it, for the same reason the laboratories do: how to dilute whisky at home sets out the method, and can you taste blending water the check that matters most.

Glossary

• Whiskey web — the weblike microstructure deposited when a microlitre droplet of diluted American whiskey evaporates on glass; composed of water-insoluble constituents that collapsed from a surface monolayer.
• Coffee-ring effect — the tendency of a drying droplet to concentrate suspended particles at its pinned edge, driven by faster evaporation at the rim (Deegan et al., 1997).
• Marangoni flow — fluid motion driven by surface-tension gradients. In whisky, differential evaporation of ethanol sets up gradients that move material along the drop’s surface.
• Monolayer collapse — the buckling and folding of a single-molecule-thick surface film when the area available to it shrinks faster than the film can relax.
• Congener — any compound in a spirit other than ethanol and water: fusel alcohols, esters, acids, aldehydes, and cask-derived extractives such as lignin degradation products.
• Micelle — a cluster of amphiphilic molecules that forms when long-chain fatty acids and esters lose solubility as water is added; the precursor of both chill haze and whiskey webs.
• Reduction (proofing) water — the purified water added to bring a spirit to bottling strength. It should contribute no solids, no flavour, and no residue.

Frequently asked questions

At what strength does a whiskey web form? Roughly 20–25% ABV. Above about 35% ABV the droplet dries to a uniform film, and near 10% ABV it dries to a coffee-ring pattern instead (Science News; Carrithers et al., ACS Nano, 2020).

Does Scotch whisky form webs? Not under the standard conditions used in the studies — Scotch and Canadian whiskies dried without the weblike structure (Science News). A Canadian and an Irish whiskey did form structures when the conditions were changed to a 2 µL droplet at 40% ABV (Physics Today, 2021).

Can whiskey webs detect counterfeit spirits? Potentially. Matching accuracy reached about 90% across 15 droplets per whiskey using image processing (Physics Today, 2021), and the authors proposed the method for authentication — but it requires a reference library of verified samples, and it complements rather than replaces isotopic and congener analysis.

Is a whiskey web the same thing as whisky “legs” or the tears of wine? No. Legs are a bulk Marangoni phenomenon — a visible film climbing the glass and running back down. A web is a dried microscopic deposit of insoluble material, only visible under magnification, although both are driven by the differential evaporation of ethanol and water.

Would mineral water change the pattern? Its dissolved solids would be left behind on the substrate when the droplet dried, adding deposit that is not part of the whiskey. That is why analytical protocols specify distilled or ultrapure water for dilution (Baxter et al., 2024; Ashmore et al., 2023) — and the same principle, at tank scale, is why distilleries proof with demineralised water rather than potable water.

Does chill filtration remove the compounds that make webs? Chill filtration is designed to remove exactly this class of material: the long-chain fatty acid esters that form micelles as strength falls (Annandale Distillery). A more heavily filtered spirit carries fewer of the colloids that a web is assembled from.

Blenders International · Water Science series