Fines: Fine for espresso, not so fine for filter by Scott Rao

Every once in a while I hear or read a misunderstanding of the role of fines in coffee extraction and flavor.  I’d like to discuss how coffee extracts, the role of fines in extraction, and how fines influence espresso and filter brewing differently.  This is a long, technical post.  But if you’re a coffee professional, it’s important you read, and probably re-read, it in its entirety.  It is easy to make practical use of the ideas I present here. 

If you find this post interesting or useful, please feel free to share, translate, or link to, this post. 


A coffee bean is a complex three-dimensional matrix of cellulose strands coated with oils and other solids.  These strands crisscross to form void spaces, or "cells."  Researchers estimate a typical bean contains a million such cells.  Roasting makes the cellulose brittle and grinding shatters the cellulose structure into various particle sizes.  We call the smallest particles “fines.”

There is some disagreement about the definition of a fine.  Most of the older research papers I’ve read referred to fines as cell-wall fragments containing no intact cells.  Some people define fines as particles with diameters smaller than a certain size, say 50 microns.  Others have recently defined fines as the particles in a certain section (the small hump, if you will) of a particle-size-distribution curve. 

 Typical particle-size distribution curve. Please ignore numbers, curve is for conceptual purposes only. Frequency here refers to weight or volume of particles.

My preferred definition of a fine is a cell-wall fragment containing no intact cells, because particles with intact cells extract differently than particles without any intact cells.  Defined this way, brewing water can freely access all of the surface of the cellulose of a fine during extraction, but in particles with intact cells, some of the cellulose surface is inside the cells and not directly exposed to the water.  

Choosing an arbitrary diameter-size cutoff to define fines makes little sense to me, as fines come in many different shapes, few of which are remotely spherical.  Unfortunately, many scientific papers make the seemingly innocuous assumption that coffee grounds are spherical, for the sake of simplifying math and modeling extraction kinetics.  That assumption is absurd and misleading, and the conclusions of such papers should be taken with a (spherical?) grain of salt.* 

Looking at the little hump of a particle-size distribution curve (PSD) is an interesting way of grouping coffee grounds, but doesn’t provide an objective criteria or definition for fines, as that hump’s quantity, quality, and distribution change along with bean, roast, grinder, burr sharpness, burr alignment, etc, not to mention the method used to measure the PSD. 


There are two phases of coffee extraction: surface erosion and inner-particle diffusion.  When water contacts a large ground-coffee particle, it rapidly dissolves coffee solids from the exposed surfaces of the particle.  In perhaps a few seconds, such erosion is nearly complete.  

Extraction from the intact cells within a ground occurs by diffusion, a process that takes minutes, and is never complete, at least under reasonable brewing circumstances.  In diffusion, a concentration gradient causes solids to move from the particle to the water (solids will always move from an area of high concentration to a neighboring area of lower concentration.)  The solids dissolved in the water then have to diffuse through the pores of the cellulose to reach the brewing liquid surrounding the particle.  Inner-particle diffusion is slow because of its many steps: water must pass through the ground’s pores, dissolve solids, pass back out the pores, and ultimately carry the solids out of the coffee slurry into the brew.  Extraction from larger particles is slower  primarily because they have longer diffusion paths. 

During a percolation process such as batch brewing, there will always be some dissolved solids left behind in the cells of the grounds, and some solids left behind in the interstitial liquid between grounds.  In an immersion brew such as a French press, extraction is slower, primarily because solids-packed brewing liquid is not continually removed from the brewer, and during most of the extraction process, the concentration gradient is not as high as it is during percolation brewing.  (This is why you should grind much finer for French press than for batch brewing, something that 99% of coffee professionals do not do.)  As well, the liquid left behind in the coffee bed of an immersion brew contains much more dissolved solids than that of a percolation brew, making immersion brewing a less efficient extraction method than percolation.   As a technical aside, the different solids concentrations remaining in the spent grounds of percolation and immersion brews is why Extract Mojo software has separate modes for calculating drip and immersion extractions.  I'm happy to write more on that in another post if readers express interest.


(and I’m sure that really excites Matt.)

Approximately 35% of the mass of a coffee bean is water soluble.  Without extreme temperatures, unrealistically long extraction times or additional solvents, it’s a fair to assume that you’ll never extract more than 35% of the weight of a pile of coffee grounds.  Quality extractions of grounds from sub-$20,000 grinders usually top out around 23% in cafes but some massive roller-mill grinders produce grounds that readily extract a few percentage points higher. 
This is speculative, but it seems that one may be able to extract about 80%—90% of the soluble solids contained on a section of cellulose (i.e., a fine) before astringency and bitterness skyrocket.  Say what?  It may be possible to get a delicious 30% extraction?  Perhaps, but not today, and not with your EK.  You may be able to get to 26% or so with a decent grinder if you grind much finer than usual, arduously sift and remove as many of the fines and “boulders” (oversized particles) as possible, and batch brew the coffee.  (Please don’t rush to judge the idea based on one such brew— the odds are low that you got the grind setting, sieving, and extraction all just right— it takes practice.  But when you do get it right, the results are fantastic.)

Dear Vince, we may need new software when we get to 30% extractions, as my CoffeeTools tops out at 26% :).

Dear Vince, we may need new software when we get to 30% extractions, as my CoffeeTools tops out at 26% :).

My contention is that if a given section of coffee cellulose is 35% soluble solids by weight, it is potentially desirable to extract 30 of that 35. (i.e. 86% of the available soluble solids), and that somewhere between 30 and 35, extraction of astringency increases dramatically.  In theory, if you had a pile of fines (fully-exposed cellulose fragments) and managed to get all of the particles to extract at exactly 30%, it would be an amazing cup of coffee; but that’s probably impossible to do. 

The problem is that in everyday brewing, the solids from the surfaces of grounds (i.e. all of the solids on the fines and exposed surfaces of larger particles) will extract beyond 30% in seconds, while the 80%? of coffee solids extracting by diffusion will take minutes to reach merely 20%.  Additionally, it’s likely that diffusion from some smaller particles or particles along channels in the coffee bed may eclipse 30% extraction before the coffee bed’s average extraction reaches 20%.   


To make the best-possible filter coffee, it’s likely you should try to minimize fines (and boulders, for what it's worth).  Given that fines are guaranteed to overextract and contribute bitterness and astringency, if you can remove fines or avoid creating them the first place, you’ll make a cleaner, sweeter cup with smoother mouthfeel.  Most of us don’t enjoy astringent coffee, but if you do, that’s great, as tannins are excellent gut-bacteria modulators.  I prefer to get my tannins from wine, thanks. 


  • Use sharper grinder burrs

  • Ensure grinder burrs are aligned as well as possible

  • Grind coarser

  • Use larger grinder burrs (burr design also matters, but larger is generally better)

  • Roast lighter

  • Avoid naturals at all costs*

  • Sift and remove the fines— this is difficult to do well, but worth experimenting with for home use and brewing competitions

  • Potentially, grind warmer coffee beans or use a warmer grinder (jury is out on this)

*Not true, but just felt like throwing that in there.

Along with minimizing fines, avoiding channeling in percolation brews is incredibly important for controlling astringency.  Fines and channels affect cup quality in similar ways. 


But what about espresso?  Espresso extraction times are too short for extraction to occur via diffusion.  Researchers have debated whether diffusion contributes at all to espresso extraction; if it does, its contribution is trivial. (I’m ignoring “espresso” made with minute-long preinfusion, "Slayer shots", and other such non-standard practices.) 

The authors of Illy’s espresso book state that fines provide the necessary extraction surface, while larger particles are required to achieve adequate flow rate through the coffee bed of an espresso.  If there are too few fines, extraction is low; if there are too few large particles, flow chokes.  Given that espresso requires rapid extraction via erosion, some fines are likely to be required for a decent shot. (I think it’s theoretically possible to isolate grounds in the 100–200 micron range and use them to pull a tasty fines-free shot, but that’s not practical in a cafe.) What’s the ideal proportion of fines?  Of course, no one knows.  To complicate the matter, there seem to be two subsets of fines; let’s call them “larger fines” and “coffee dust.”  The presence of the tiniest fines, or coffee dust, probably damages espresso quality by disproportionately slowing flow rates relative to their extraction contribution, thus requiring a coarser grind setting, which in turn decreases extraction. 

Chris Hendon noted on my Instagram recently that grinding frozen beans produces less coffee dust and more of the larger fines, and they are of more uniform size.  I have not seen the data, but that phenomenon may explain the success many baristas claim to have with grinding frozen beans for espresso.  Minimizing coffee dust and producing larger, more uniform fines would almost certainly produce superior espresso extractions.  Thanks, Chris, and I hope I have represented your findings accurately.


Congratulations if you’ve made it through this whole post.  Now please read it again so it all makes sense :).  Most importantly, I hope this post convinces you to be cognizant of fines production and the ways you can minimize it when grinding for non-espresso brewing.  The two most impactful things most cafes can do to immediately improve quality are to replace old, worn grinder burrs with fresh, seasoned burrs, and to align the burrs as precisely as possible.  Extractions will increase, flavor clarity will improve, astringency and bitterness will decrease, and mouthfeel will be smoother. 

I’d love to know your thoughts.  Don't be shy. 

Thank you to Scott Rao for allowing us to use this article. For great products and more articles like this, please visit:

Leave a comment

Please note, comments must be approved before they are published