Grind size explained

A grinder makes a distribution

Sourced

When researchers measured the particles from a large shop grinder, counted by number, they found two populations at once: a great many very small particles, and far fewer large ones ten to a hundred times bigger across. The small ones, which brewers call fines, accounted for most of the surface area available to water. The large ones mattered even in small numbers.

The same study compared four coffees of different origin and processing, all roasted light to medium, at one setting on that grinder. The distributions were very similar. The authors concluded that, for coffees like these, the spread a grinder produces does not depend on where the coffee came from or how it was processed, and added that they could only speculate about darker roasts.

They also wrote that it is not clear whether an ideal particle size distribution exists. That is worth keeping in mind when a grinder is sold on the uniformity of its output.

What fineness does to a brew

Widely documented

Finer coffee offers more surface and shorter distances for dissolved material to travel, so it gives up more in a given time. In any brewer where water passes through a bed, finer coffee also resists the flow, so the water stays longer. The two effects push the same way, which is why a small change of grind moves a pour-over or an espresso so much.

The trend has a limit. In one study of espresso, extraction rose as the grind became finer and then fell at the finest settings, which the authors attributed to the bed clogging in places so that water bypassed part of it. A second study, on a different grinder and machine, saw the same rise and fall in what reached the cup, and found that grind settings so close together that their measured distributions were nearly identical still produced different pressures in the bed. Small differences in the fine fraction appear to matter more than the setting's number suggests.

Colder beans grind differently

Sourced

The first study also ground one coffee at four bean temperatures, from room temperature down to that of liquid nitrogen. Colder beans gave a narrower distribution and a smaller average particle. The measured cut-off for the fine fraction fell from about 70 micrometres at room temperature to about 61 at the coldest.

The freezer result depends on which measure is read. Its cut-off for the fine fraction, 73 micrometres, was within the stated error of the room-temperature figure. The spread, though, narrowed most between room temperature and the freezer, according to the authors, and the most common particle size fell steadily with cooling, by 31 per cent at the coldest. The authors brewed and tasted nothing, so the study shows that bean temperature changes the grind and stops short of showing what that does to a cup.

Static, clumps and a few drops of water

Sourced

Grinding charges coffee with static electricity, which makes it clump and cling to the grinder. A study that compared ways of reducing the charge found that adding a small amount of water to the whole beans, under 0.05 millilitres per gram, cut the charge by at least half, stopped clumps forming and kept fine particles from sticking inside the grinder. Espresso made from a dark roast ground this way ran longer and was at least 15 per cent more concentrated. A light roast of the same coffee showed no significant change.

The authors used one large grinder and three coffees and say other coffees may need different amounts of water or behave differently. They measured concentration and did not taste. They also note that too much water may cause caking or corrosion, with a limit that depends on the coffee and the grinder, so the maker's instructions for a given grinder come first.

How a distribution is measured, and why the method changes the answer

Sourced

Three kinds of instrument appear in the papers read for this page, and they do not measure the same thing. Laser diffraction passes a beam through a cloud of particles and works the sizes back from how the light scatters. The grinding study used it and took particular care over the alignment of the optics, because ground coffee spans three orders of magnitude in size and the largest particles are hard for the technique. Image analysis photographs falling particles and measures the area each one presents to the camera; a Swiss study of espresso used that, and reported its median as a volume-weighted figure worked out from those projected areas. Sieving shakes the grounds through a stack of meshes and weighs what each one holds; a thesis on gas in coffee used it to define coarse, medium and fine, and the Swiss group used a single 120 micrometre sieve to collect fines.

None of these gives the size of a coffee particle in any plain sense, because the particles are not spheres. Each reports the size of the sphere that would behave the same way in that instrument: scatter light alike, cast the same shadow, or just pass the same hole. A flat flake and a round grain of equal weight come out differently in each. A figure in micrometres belongs to the instrument that produced it as much as to the coffee.

The second choice is how to count. The same sample can be described by number of particles, by their surface or by their volume, and the three pictures barely resemble each other. In one study's coffee at one espresso setting, 99 per cent of the particles by count were under a tenth of a millimetre, and they carried 80 per cent of the surface; yet a chart of the same sample by volume is dominated by the far fewer large particles, because one particle ten times wider holds a thousand times the material. A median by volume says where the bulk of the coffee's weight sits. It says almost nothing about the fines.

That is why the Swiss group argues against the single number. In earlier work on capsules, they write, collapsing a full distribution into one figure such as the median did not describe how the coffee extracted well enough, and a second figure, the share of the volume in particles under a tenth of a millimetre, did more to explain shot time than the average size did. Two grinders can share a median and differ in that share.

A model for reading any particle-size chart: four questions

CoffeeHQ explanation

A model, in the sense of a checklist. A particle-size chart published by a grinder maker, a reviewer or a hobbyist with a phone and an app can be placed by asking four things of it. Without the answers, two charts cannot be compared.

Counted how?

By number, by surface or by volume. A by-number chart of almost any coffee is a spike of fines; a by-volume chart of the same coffee is a hump of large particles with a small shoulder. Neither is wrong and they cannot be laid over one another.

Measured how?

Diffraction, imaging or sieves, each with its own idea of size. A phone camera cannot resolve the smallest particles, so a chart made that way under-reports fines by construction.

What was left out?

Grounds that stayed in the grinder, clung to the chute by static or were lost in handling are not in the sample. Static holds back the finest particles preferentially, on the evidence of the study that reduced it with a little water.

Which coffee, at what temperature?

In the one study that varied it, colder beans gave a narrower spread and a finer most-common size. The same grinder at the same setting gives a different chart from a different roast or a bag out of the freezer.

Burrs: where the claims outrun the published data

Accounts differ

A great deal is said about burrs. Flat burrs are said to give a narrower spread and more clarity, conical burrs more fines and more body; larger burrs, slower burrs, coated burrs, burrs that have been seasoned by grinding several kilograms of coffee and burrs aligned to within a hair are each credited with a better cup. CoffeeHQ searched three indexes of research literature for a peer-reviewed comparison of any of these and found none: no study of flat against conical burrs, of burr speed, of alignment, of wear, of seasoning or of how fast beans are fed in. The nearest thing found is a 2026 manuscript by one author, not yet peer-reviewed and read only in its summary, which photographed grounds from one conical and one flat grinder at six settings each; it asks which summary of a distribution best predicts the strength of the brew, not which burr is better. Enthusiasts and trade writers have published their own measurements of grinders online, and those are not drawn on here. That is a finding about the research literature, not about grinders. Makers certainly hold test data, and some of these effects may be real.

What the research papers do contain is incidental, and worth having because it is all there is. The grinding study notes that all burrs are liable to become misaligned, meaning the two discs are not perfectly parallel; that alignment can be judged roughly by the pitch of the sound the burrs make as they are brought together; that a distribution's smoothness and spread also carry information about alignment, though adjusting by that is slow and difficult; and it shows one distribution from a misaligned set in its supplement. Its authors had three grinders of the same model to choose from, two with one burr pattern and one with another, and chose between them by which was making sellable espresso that morning. That is a candid picture of how much is controlled.

The Swiss group states, citing its own earlier work, that for a given median size different grinding technologies give different shares of fines. That is the nearest thing to support for the idea that burr design matters, and it is a statement about fines, not about flavour. The same group's direct test found no penalty in one taster's scores for a higher share of fines.

So the honest position on burrs is narrow. Design affects the share of fines for a given median; alignment affects the spread; nobody has published what either does to taste. A claim that a burr shape gives clarity or body is a trade expectation, and a reader's own comparison is better evidence than any description of one.

Why a setting does not travel

CoffeeHQ explanation

A number on a grinder's dial is a position of its burrs, in units the maker chose. Two grinders at the same number produce different spreads, and the same grinder produces a different spread as its burrs wear or after it is taken apart and reassembled. The studies above each report their setting as an arbitrary figure on one machine for this reason.

It follows that a recipe's grind instruction is a starting point at best. What transfers between kitchens is the result to aim for: how long the brew takes, how it flows, and how it tastes. A size in micrometres does not travel either, for the reasons given above: it belongs to an instrument and a way of counting.

Where each method usually starts

Typical range

• Espresso — The finest of the common grinds, because the water has seconds and is forced through; the range that works is narrow and specific to the grinder.
• Pour-over — A middling grind, adjusted so the water drains in a few minutes; flow time is the practical guide to whether it is right.
• French press — Usually coarse, mainly so the mesh can hold the grounds back; in immersion the grind has less effect on where extraction ends up.

What this page does not cover

CoffeeHQ explanation

It gives no size in micrometres for any brewing method, since the figures above describe one grinder at one setting. It makes no claim about what conical or flat burrs, burr speed, alignment or seasoning do to a cup, because no peer-reviewed comparison of any of them was found. It does not say a narrower spread tastes better; the study that measured spreads says that is unknown. And it does not recommend adding water to beans, which is a decision for the grinder's maker and its owner.

What to do next

• Grind-size guide — Places your method on a scale of named steps and separates flow problems from taste problems.
• Grinder burrs — What burrs are, how they wear, and a comparison of the two shapes labelled as a model.
• Choosing a grinder — Which kind of grinder suits which kind of brewing, without brand recommendations.
• Coffee extraction — What the grind is being adjusted to achieve.