Coffee facts · Brewing coffee
Grinding and extraction in batch brew: what happens after the bean is ground
Short answer: Grinding does two things to coffee. It shortens the diffusion path for gas and aroma compounds from millimetres to tenths of a millimetre, and it creates a particle size distribution that governs extraction. Up to 75 per cent of the gas in freshly roasted coffee is released within 90 seconds of grinding, and after that oxygen gets in. A wide particle distribution gives over- and under-extraction at the same time, and poorer repeatability. Batch brew is the least forgiving method for both.
1. The gas: measured, not assumed
Roasting forms carbon dioxide, and much of it stays trapped in the bean’s pore structure. How quickly it leaves the coffee depends almost entirely on whether the bean is whole or ground.
- Whole bean: more than 800 hours, just over 33 days, to release around 90 per cent of the remaining carbon dioxide (Wang & Lim, 2014).
- On grinding: coarse, medium and fine grinds lost around 26–30, 33–38 and 45–59 per cent of the remaining carbon dioxide respectively (Wang & Lim, 2014).
- Time-resolved: using a gravimetric method that tracks the loss of mass, Smrke et al. (2017) found that up to 75 per cent of the trapped gas in freshly roasted coffee was released during and immediately after grinding, within 90 seconds, up to the first measurement point.
Two mechanisms are behind this. Grinding opens pores and cells directly, which explains the instant loss. The rest is governed by diffusion out of the particle.
The instant loss happens wherever the coffee is ground. It is not a deadline for brewing, and some of the gas you actually want gone: too much carbon dioxide disturbs the wetting of the coffee bed. That is why roasters let coffee rest for a few days after roasting. What separates the grinding options is what happens after the first few seconds.
2. Why particle size dominates: t ∝ L²
Diffusion time out of a body scales with the square of the characteristic distance: t ≈ L²/D, where D is the diffusion coefficient through the material.
| Characteristic distance L | Relative diffusion time | |
|---|---|---|
| Whole bean, centre to surface | approx. 2 mm | 1 |
| Filter-ground particle, centre to surface | approx. 0.3–0.5 mm | 1/16 to 1/50 |
| Fines | a fraction of that | further orders of magnitude shorter |
This is a simplification. The bean is not homogeneous, D changes as the gas thins out, and the distribution contains every size at once. But the scaling explains why the difference between whole and ground is not a matter of per cent but of factors, and why a finer grind loses more: the distance is squared.
The same reasoning applies to the volatile aroma compounds, which travel partly with the gas and partly sit in the oils of the cell walls. Once the gas is gone, oxygen gets in. Oxygen is the dominant driver of degradation after roasting. According to the SCA, lowering the oxygen level from the 21 per cent in air to 0.5 per cent can extend shelf life up to twenty times, and ground coffee ages several times faster than whole beans. The most aromatic sulphur compounds are among the least stable. In storage studies the freshness marker methanethiol fell to around 30 per cent of its original amount after eight days (Holscher & Steinhart, 1992).
3. From grinder to brewer: two timelines
| Step | Pre-ground, vacuum-packed | Whole beans, ground at the brewer |
|---|---|---|
| Grinding | At the roastery | At the brewer |
| Before packing | The coffee has to degas, or the vacuum pack will inflate | The beans are packed whole in a valve bag |
| After opening | Manufacturers state the pack should be used within 10–14 days | The beans stay whole until the dose is ground |
| Time from grinding to brewing | Weeks to months | Under a minute |
The vacuum pack protects the coffee that remains from further oxidation. But the instant fraction and the protective gas are gone before the pack is sealed, by design. Coffee ground in the morning for the whole day ends up in between, with hours in the diffusion phase.
4. The particle distribution: the grinder, not the bean
Roasted coffee is brittle and fractures; it is not cut. A grinder therefore always produces a distribution, typically bimodal: a peak of fines from cell wall fragments and a main peak around the set grind size.
Uman, Hendon et al. (2016) showed that the distribution was largely independent of the bean’s origin and processing method. It is the grinder’s geometry and setting that determine it.
Extraction from a particle follows the same scaling as the gas. Fines extract far faster than the main fraction, which in turn extracts faster than the coarsest particles. A wide distribution therefore gives, at the same time:
- over-extracted fines, with bitterness and dryness,
- under-extracted coarse particles, with sourness and thinness.
The sum can have an acceptable average and still taste flat. Adjusting the grind size moves the whole distribution, not its width. You cannot brew your way out of it.
The distribution also affects flow. Cameron, Hendon et al. (2020) showed for espresso that fines and uneven packing cause channelling, lower extraction yield and poorer repeatability. The mechanism, water taking the path of least resistance through a bed of mixed particle sizes, applies to percolation in general and therefore to filter brewing.
5. The target: extraction yield and strength
Extraction yield (EY) is the share of the coffee’s dry mass that ends up in the cup:
EY = TDS × beverage mass ÷ dose
The Golden Cup standard specifies TDS 1.15–1.35 per cent and EY 18–22 per cent. The ranges come from consumer studies in the 1950s and are not a law of nature, but they are the common reference that refractometers and brewers are calibrated against.
The window is narrow enough to demand two things: a distribution narrow enough that EY means something for each particle and not just as an average, and a dose that is the same every time.
6. Dose sensitivity, honestly calculated
With a constant amount of water and constant EY, TDS is roughly proportional to the dose.
| Dose | Deviation ±2 g | TDS at a 1.25 % target |
|---|---|---|
| 60 g | ±3.3 % | 1.21–1.29 % |
| 125 g | ±1.6 % | 1.23–1.27 % |
| 225 g | ±0.9 % | 1.24–1.26 % |
A two-gram deviation stays within the Golden Cup window. That is not where the problem lies.
The problem is that a scoop measures volume, not mass. The density of the beans and the ground coffee varies with roast degree, grind size and how the scoop is filled. Changing coffee or roast degree therefore moves the dose without anyone noticing, and the deviation is systematic, not random. Weighing removes that variable entirely.
7. Batch size and bed depth
A larger dose in the same filter basket gives a deeper coffee bed, longer contact time and a different flow than a small one. The same grind size therefore does not give the same EY at 60 g as at 225 g. Anyone brewing several batch sizes in the same brewer needs one setting per batch size, or has to accept that one of the batches lands outside the target.
8. Why batch brew is the least forgiving
Espresso has high pressure, short contact time and a concentration several times higher, which masks a great deal. Batch brew has just over one per cent dissolved solids. Every deviation in distribution or dose shows immediately, and a bad airpot is a mistake in many cups. It is also the method where most businesses have the least process control: a scoop and pre-ground coffee.
9. Technical data: Science Brewing Grinder
For anyone who wants to translate the above into a specification:
- Burrs
- 50 mm, flat, hardened steel, geometry for filter grinding
- Adjustment
- Stepless, micrometric. The recommended setting is shown on the screen; the burrs are set by hand
- Clumping and static
- Anti-static system
- Drive
- Direct drive, 1,350 rpm
- Grind speed
- 1.5–2.1 g/s. 125 g ≈ 60–85 s
- Scale
- Resolution 0.1 g
- Dosing, brew doses
- Fast mode, stops within a couple of grams of the target
- Dosing, small doses
- Precision mode with short pulses, within a few tenths of a gram
- Freshness in the grinder
- Discards a small amount if the coffee has been standing longer than the chosen time (30 min–48 h). Can be switched off
- Log
- Time, selection, target weight and actual weight per grind. Syncs after an interruption
- Stock
- Every grind is deducted from the stock of roasted coffee, with alerts
- Power
- 230 V single phase, 350 W, Wi-Fi
- Price
- 637 € excluding VAT (grinder and base platform). Base platform on its own 364 € excluding VAT
- Delivery
- 2–4 weeks. Built and checked to order in Sweden
The log is what makes it useful for anyone who measures. Actual versus target weight per grind, linked to which coffee is being ground, is the same data you would otherwise keep by hand. For roasteries that place the grinder with café customers, consumption is visible per customer: Grinder for roasteries.
If you roast yourselves, the roasted coffee from Science Brewing Roaster goes into the same stock the grinder draws from. Then you know the roast date and resting time of every dose in the grinder.
Read more
- For the plain-language version: Whole beans or pre-ground? The clock starts when the coffee is ground
- For judging the coffee in the bag: How can you tell if coffee is fresh?
Sources
- Wang, X. & Lim, L.-T. (2014). Effect of roasting conditions on carbon dioxide degassing behavior in coffee. Food Research International.
- Smrke, S., Wellinger, M., Suzuki, T., Balsiger, F., Opitz, S. E. W. & Yeretzian, C. (2017). Time-resolved gravimetric method to assess degassing of roasted coffee. Journal of Agricultural and Food Chemistry.
- Uman, E., Hendon, C. H. et al. (2016). The effect of bean origin and temperature on grinding roasted coffee. Scientific Reports 6:24483.
- Cameron, M. I., Hendon, C. H. et al. (2020). Systematically improving espresso: insights from mathematical modeling and experiment. Matter.
- Holscher, W. & Steinhart, H. (1992), on methanethiol during storage, cited via the SCA.
- Specialty Coffee Association, Coffee Freshness Handbook, and the Golden Cup Standard.
Frequently asked questions
How fast does ground coffee release carbon dioxide?
Up to 75 per cent of the trapped gas in freshly roasted coffee is released within 90 seconds of grinding. Whole beans need more than 800 hours to release around 90 per cent of the remaining gas.
Does the coffee have to be brewed within 90 seconds?
No. The instant loss happens wherever the coffee is ground. What matters is the time after that, when oxygen gets in: minutes at the brewer, weeks to months for pre-ground.
Why does a finer grind lose more?
Diffusion time scales with the square of the distance out of the particle. Halving the particle size gives roughly a quarter of the time.
Can I compensate for a wide distribution with the grind size?
No. The grind size moves the distribution, not its width. Fines and coarse particles still extract at different speeds.
Do two grams matter?
For 125 grams, two grams shift the strength by about 1.6 per cent, which is within the Golden Cup window. What matters is that a scoop measures volume and that density changes with roast degree and coffee.
Do different batch sizes need different grinds?
Often, yes. A larger batch gives a deeper bed and longer contact time. The same setting rarely gives the same extraction at 60 and 225 grams.
Who we are
Science Brewing exists to end bad coffee. We sell green coffee and build Science Brewing Roaster and Science Brewing Grinder. We roast our own production on the machine, in Halmstad, Sweden.