Initially posted: 01 Jun 16
This study had two goals: 1) explore the impact of uniform particle size on espresso extraction; 2) explore the impact of particle stratification on espresso extraction. The dependent variables collected were shot time and TDS. Coffee particles were stratified into two ranges: 177-250 µm and 250-300 µm. Six conditions were tested with shots consisting of: 1) particles between 250-300 µm layered in the bottom half of the portafilter basket and particles 177-250 µm layered on the top; 2) particles between 177-250 µm layered in the bottom half of the portafilter basket and particles 250-300 µm layered on the top; 3) particles solely of 177-250 µm size; 4) particles solely of 250-300 µm size; 5) thoroughly mixed particles of 177-300 µm size; 6) standard shots (no particle manipulation). Results demonstrated significant effects of condition on TDS and shot time (p = 0.00).
Based on our data, given a standard espresso setup, particle size alone did not systematically predict espresso extraction (TDS/extraction yield). A larger range of particle size variability led to increased extraction. Further, particle size did significantly impact overall shot time (i.e., smaller particles led to longer extraction time); however, the placement of smaller particles within an espresso puck did not lead to significant differences in shot time or extraction.
We have no vested interest in any of the products being used for this experiment. STANDART was given exclusive initial publication of the write-up for this experiment (StandArt, 4th Edition, February 2016).
Jeremy, Joe, and Dave
The impact of particle size and placement within an espresso puck is an active area of interest and debate for the coffee community. There are many high-end academic and industry labs investigating nuances related to the dynamics of fluids through an espresso puck (e.g., Corrochano et al., 2015). Our approach differs, however, as we asked ourselves two relatively simple questions and attempted to investigate them in a very applied, practical way: does the position of particles within a puck affect its extraction and does the uniformity of particle size affect espresso extraction.
Within the domains of basic and applied science, there is always a balance between internal and external validity. So while a more sophisticated approach can create an experimental environment of strong controls, potentially yielding detailed information on underlying mechanisms, the practicality and relatability of its methodology would be foreign to a coffee shop, barista, or consumer. Our approach is targeted on the applicability to a shop or individual, placing significantly greater value on the external validity of our findings.
The two main hypotheses were 1) controlled placement of particle sizes would lead to significant differences in extraction (“fines migration” hypothesis) and 2) more uniform particles would lead to significant differences in extraction (“unimodal distribution” hypothesis).
The coffee used for the experiment was a single origin washed Ethiopian Sidamo Guji. Coffee was rested for 12 days before use. The same batch of roasted coffee was used for the entire experiment. Room temperature was 25°C and humidity 65%.
- La Marzocco (LM) GS/3 MP (mechanical paddle), single group fitted with 0.6mm restrictor and standard double portafilter with the 20g VST basket filter, set at 9 bar water pressure (verified with a Scace II); brewing temperature 94°C regulated with a PID on board
- Water filter used for the experiment was Brita Purity C150 Quell ST
- Mahlkönig EK43 coffee grinder (the same grind setting was used for all conditions)
- Two scales (Ohaus Navigator XL calibrated prior to the experiment to measure the dry coffee dose; Acaia Lunar calibrated prior to the experiment to measure the beverage mass)
- HG1 blind doser
- Pen and paper to record values
- 5 (for each condition) empty ramequin bowl for measuring the mass of the shot
- 5 (for each condition) ceramic/glass cups for TDS measurement
- Eazytamp 5 Star Pro tamper to standardize the levelness and axial compression of the coffee bed (verified at 3 kg of pressure)
- VST LAB III 4th Generation refractometer, zeroed according to manufacturer guidelines
- Distilled water
- Alcohol pads
- Infra-red gun to measure the solution prior to refractometer measurement (under 30°C degrees, average 27.5°C)
- Air Blower/Super Soft Microfiber towels
- Advantech Manufacturing DuraTap & 8” Half-Height Stainless Steel Test Sieves, consisting of:
- US Std. Mesh Size No. 50 Sieve (300 µm)
- US Std. Mesh Size No. 60 Sieve (250 µm)
- US Std. Mesh Size No. 80 Sieve (177 µm)
- Bottom Pan
Figure 1. Sieve stack used for experiment.
The experiment was carried out over 8 hours. The sieve stack was assembled and loaded with 100 g of coffee ground from the EK43 in the No. 50 sieve. The DuraTap was then used to perform controlled shaking and tapping for 10 minutes, per the “optimized stack” protocol established by Socratic Coffee previously. This was repeated as needed to arrive at amounts of coffee required in the desired particle size ranges.
Six different particle conditions were tested, with 5 shots per condition.
Condition 1 – Layered puck with 10 g of particles in size range 177-250 µm on top and 10 g of particles in size range 250-300 µm on bottom.
Figure 2. Condition 1 with the smaller particles placed on top of the puck.
Condition 2 – Layered with 10 g of particles in size range 250-300 µm on top and 10 g of particles in size range 177-250 µm on bottom.
Figure 3. Condition 2 with the larger particles placed on top of the puck.
Condition 3 – 20 g of particles in size range 177-250 µm.
Figure 4. Condition 3 with uniform particles between 177-250 µm.
Condition 4 – 20 g of particles in size range 250-300 µm.
Figure 5. Condition 4 with uniform particles between 250-300 µm.
Condition 5 – 10 g of particles in size range 77-250 µm and 10 g of particles in size range 250-300 µm, all thoroughly mixed.
Figure 6. Condition 5 with two equal parts of stratified particles thoroughly mixed.
Condition 6 – Control condition with standard 20 g dose shots (no particle size manipulation).
Figure 7. Condition 6 with completely unsifted particles.
The conditions with the grinder and espresso machine were constant for all conditions. The room temperature was controlled with air conditioning at approximately 25 °C for the duration of the experiment.
A brew ratio of 1g coffee to 2g brew weight was used (i.e., 20g dose for 40g final beverage mass). Because TDS is most strongly correlated with beverage mass, shots were all pulled to a consistent mass. Time to reach this weight was recorded. Shots were performed with 10 s pre-infusion before activating pump.
All TDS samples were performed following VST recommended guidelines.
All analyses were performed using R 3.1.1 and Excel 2016. Data was first assessed to ensure it does not violate assumptions for multivariate analysis of variance (MANOVA). Beverage mass values between conditions were compared via t-tests. No significant beverage mass differences existed between conditions with the exception of the control condition. Variance in beverage mass for that condition was found to be much lower than the others. A MANOVA utilizing the Pillai test, with shot time and TDS as dependent variables, showed a significant effect of condition (overall: F(5,24)=4.76, p=0.00; shot time: F(5,24)=26.12, p=0.00; TDS: F(5,24)=2.95, p=0.03). More specific ANOVAs and t-tests were then utilized to better explore relationships in the data.
The below plot shows TDS values and shot times, clustered by condition.
Figure 8. Scatterplot of TDS values and shot times to reach 40 g beverage mass.
Extraction yield can be calculated simply as: (Beverage Mass x TDS)/Coffee Dose
Because our experiment controlled coffee dose and aimed for consistent beverage mass for all conditions, using extraction yield as a dependent variable did not yield significantly different results compared to using TDS (see Figure 9). As such, we completed all further analyses only with TDS.
Figure 9. Scatterplot of extraction yield values and shot times to reach 40 g beverage mass.
Comparing the effects of particle layering (Conditions 1,2, & 5) on TDS (Figure 10) and shot time (Figure 11), demonstrated no significant effects (p=0.39 and p=0.66, respectively).
Figure 10. Boxplots showing the average with variability of TDS for shots in each condition.
Comparing the effects of particle uniformity and size (Conditions 3,4, & 6) on TDS (Figure 10) and shot time (Figure 11), demonstrated a marginally significant effect, F(2,12)=3.71 (p=0.06), for TDS and a significant effect, F(2,12)=14.85 (p=0.00), for shot time.
Figure 11. Boxplots showing the average and variability of time for shots in each condition.
Looking more closely at the effects of particle uniformity and size, contrasts between conditions revealed significantly lower TDS from particles of 177-250 µm compared to 250-300 µm and control (p=0.00), but not compared to the 50/50 mix of the two particle stratifications (p=0.08). As for total shot time, all perturbations of comparison between conditions differed significantly (p=0.00 for all)—i.e., 177-250 µm only differed from all other conditions (control, 50/50 mixed, 250-300 µm only); 250-300 µm only differed from all other conditions (control, 50/50 mixed, 177-250 µm only); 50/50 mixed differed from all other conditions (control, 177-250 µm only, 250-300 µm only); and, the control differed from all other conditions (50/50 mixed, 177-250 µm only, 250-300 µm only).
(Raw data can be downloaded in a tab delimited text file here. As always, while we offer the data for your personal use, we kindly ask that you send a message to email@example.com before posting or presenting it in any public forum and attach appropriate acknowledgement.)
It is useful to think of this as two mini-experiments (Table 1), each with two dependent variables. One experiment tested the placement of particle sizes within the puck, quantifying the effect on TDS and shot time (Conditions 1, 2, 5, & 6). In some ways, this can be thought of as looking at the oft-mentioned concept of “fines migration”. The other experiment tested the effect of particle size alone on TDS and shot time (Conditions 3, 4, 5, & 6). This can be thought of as looking at uniformity of grinds for espresso extraction. Conditions 5 & 6 serve as control conditions for comparison. Though the control condition incorporated more small particles into the puck, it incorporated far more large particles (Figure 12).
Figure 12. Particle size distribution of the control condition EK43 shot.
Table 1. Summary of the results from the experiment, showing the conditions compared for each “mini experiment”.
(Conditions 1, 2, 5, & 6)
|Uniform Particle Size
(Conditions 3, 4, 5, & 6)
|Impact on TDS||Placement of particles did not lead to significant differences.||Impact on TDS||Uniformity of particles did have a significant effect. Smaller particles led to less TDS.|
|Impact on Shot Time||Placement of particles did not lead to significant differences.||Impact on Shot Time||Uniformity of particles did have a significant effect. Smaller particles led to longer shot times.|
Tortuosity, or the diffusion of water through the coffee bed, is the main principle we are most likely affecting here. The physics of the interactions between particle size, morphology, roast development, axial compression and the coffee bed matrix, etc. requires significant and complex methods to examine each variable in isolation. But we are often left wondering how to apply the knowledge gained to a café or home setting. For instance, the often discussed “fines particles migration” was described by Petracco and Liverani (1993) as they observed “transient flow” changes upon reversing the direction of the water through a packed coffee bed. Even the earlier cited study (Corrochano et al., 2015), which discusses many of the variables impacting flow through a coffee puck, assessed the permeability and flow rate of an espresso puck after extracting the coffee for 600 s. Reversed flow and 10 minute extractions are not typical approaches to coffee brewing. We argue that our methodology here is much more applicable to an actual café or home user setting.
While we are the first to admit our work is not absolutely conclusive, it provides well-controlled, objective data addressing two commonly discussed concepts in coffee: “fines migration” and “particle distribution for espresso”. We would argue, based on our data, that the theory of fines migration does not appear to play a critical role in espresso extraction with respect to TDS and shot time, as performed by the average consumer. Further, we would argue that uniformity of particle size for espresso does not seem to yield beneficial effects if your goals are rapid and thorough extraction of coffee solubles. Unexpectedly, smaller particles led to less overall extraction but greatly increased shot time. We suspect this is due to a phenomenon often referred to as “jamming” in the study of fluid flow through porous media. This jamming effect may lead to increased resistance faced by the water, hence the increased shot time, but also blocking of particle surfaces for potential extraction. Our work does not address what level of variability might be ideal for espresso extraction; it simply suggests some variability leads to faster and more efficient extractions.
We intentionally did not collect or present any subjective data. We feel it is the role of the barista/café owner to determine the desired taste given his/her equipment, coffee, water, and preference. Our data is meant to compliment and better inform the coffee community so that they may arrive at their desired taste more efficiently, effectively, and consistently.
Corrochano BR, Melrose JR, Bentley AC, Fryer PJ, Bakalis S. 2015. A new methodology to estimate steady-state permeability of roast and ground coffee in packed beds. J Food Eng, 150, 106-116.
Petracco M, Liverani FS. 1993. Dynamics of fluid percolation through a bed of particles subject to physio-chemical evolution, and its mathematical modelization. In: ASIC (Ed.), 15th International Conference on Coffee Science, 702-711.