Why are conical burrs inherently bimodal? - Page 2

Grinders are one of the keys to exceptional espresso. Discuss them here.

#11: Post by kmills »

Below are two resources worth a read for anyone interested in particle size characterization. The take away is that the measurement technique (laser scattering) is not well suited to coffee's irregular shape and the analysis of the data is complex when considering both number and volume distributions. The graph above is Volume distribution.

Fines are both a contributor to the flavor (via extraction rate) and a contributor to the kinetics of the extraction via water flow restriction; without sieving, you cant decouple the parameters. When you look at a number distribution, the fines are better represented and point to why overextraction can occur. One can not easily independently test the impact of fine size and count without impacting coarse size and count AND the extraction its self.

Immersion brewing doesn't suffer the convolved variables, I believe it is well understood that the presence of excess fines in brewed coffee is easily detectable.

http://www.brookhaveninstruments.com/pd ... 202011.pdf
http://www.horiba.com/fileadmin/uploads ... debook.pdf

User avatar

#12: Post by Terranova »

kmills wrote:The take away is that the measurement technique (laser scattering) is not well suited to coffee's iregular shape and the analysis of the data is complex when considering both number and volume distributions.
It depends which clues you are after, agreed there is not such a thing which can accurate measure the particle size because of there irregular shape, but to deal with these difficulties, definitions of 'statistical geometric diameters' were established.
An example is Martin's diameter, which is the length of the chord that divides the cross-sectional shape into two equal areas. Another is Feret's diameter, which is the distance between two parallel lines tangent to the projected cross-section.


#13: Post by kmills »

Correct, figure 1 in the Horiba link introduces this concept. Laser scattering automatically and necessarily assumes a sphere, however. A mono-disperse population of oblong shapes will show a broad peak (or even bimodal) due to this; we found that this was a confounding aspect of measuring coffee and prevented meaningful conclusions. Automated optical techniques are requred for further analysis.

I run into this issue when comparing metal powder produced by gas atomization (spheres) and water atomization (complex organic shapes) which are nominally spec'd to the same sieve size. Coincidentally, they behave very differently in a paint-like slurries due to their differing morphology despite the same "size".

User avatar
Team HB

#14: Post by another_jim »

The laser distributions are by surface area. The distributions where fines are highest are by particle count. Summing over the volume shows fines are always less than 1/2 percent of the coffee by weight. Again, becuse of the faulty shape approximations, these figures are tentative.

What isn't tentative is tasting various sieved coffees steeped then filtered. I for one cannot tell whether fines were involved or not. As a first approximation, fines regulate flow in percolation methods, they affect mouthfeel in unfiltered brewing, but they do not directly affect taste.

However, one aspect is worth exploring. Fines, and smaller particles in general, should absorb brewing water faster than coarser particles. This means that in percolation methods, where the ground coffee might not reach an absorbtive equilibrium, they may have some systematic under extraction effect on beyond their regulation of flow. That is, you are using the fine ground coffee in a pour over. The water takes longer to get through the bed, so you pour more slowly. The early water is preferentially absorbed and held by the fine particles; the coarse particles are starved until later in the pour, and never fully extract.
Jim Schulman