The Chemistry of Pre-Infusion: What Happens in the Puck in the First Ten Seconds
The Chemistry of Pre-Infusion: What Happens in the Puck in the First Ten Seconds
"10 seconds of pre-infusion double the body" – you read that sentence a lot. We went looking for the source and found none. What does exist, since 2025, is something far more interesting: X-ray images of what actually happens inside the puck.
The water front needs 5 to 10 seconds to work its way through the puck and displace the trapped air along with the residual CO2. That is measured with X-ray microtomography.
In the same window the particles swell by around 15 percent. The effective pore diameter drops from about 200 to 6 micrometres, the permeability by roughly a factor of a thousand. So the flow resistance comes mostly from swelling, not from compaction.
That pre-infusion doubles the body is not documented anywhere. The only measurement series we know of on pre-infusion time and mouthfeel even finds the opposite trend. We are writing that down, even though it ruins the nicer story.
The water front: 5 to 10 seconds
Until recently the opening phase was a genuine research gap. Writing in Matter in 2020, Cameron and Hendon state explicitly that their espresso model only becomes valid after liquid infiltration has taken place. So the most-cited espresso model says itself that it has nothing to say about pre-infusion.
In 2025 Foster and colleagues filmed it directly for the first time, with a rotating X-ray system that tracks the infiltration front through density contrast. In 2026 a paper by Waszkiewicz and colleagues followed, using X-ray microtomography at 20 micrometre resolution, pressure sensors and a poroelastic model. Their description of the first seconds is unusually vivid: water enters a complex pore network saturated with air and residual carbon dioxide, and first has to push that gas out.
This is exactly the window in which it is decided whether the bed gets evenly saturated. If it does not, the water then looks for the fastest route under full pressure: channeling.
Swelling, not compaction
The second finding corrects the usual barista story. It is often said that the puck is squeezed together by pump pressure and becomes denser as a result. Waszkiewicz et al. measure something else: roasted coffee particles take up water very quickly, and the polysaccharide-rich cell structure expands hygroscopically. Clamped inside the basket, this swelling reduces the free pore space.
That also explains why a shot with a long pre-infusion runs more slowly later on: the puck has more time to swell before full pressure arrives.
Peer-reviewedThe same paper delivers two more numbers worth remembering: the pressure at which Darcy behaviour ends and the flow saturates sits at 12 ± 3 bar. And the TDS decay time within the shot is 8.8 ± 1.5 seconds, so most of what gets dissolved dissolves early.
Typical pressures and durations
Manufacturers distinguish two things that often get mixed up. La Marzocco defines pre-brew as briefly wetting the puck surface in an off-on sequence, and pre-infusion as continuous saturation at low pressure before pump pressure rises.
| System | Pressure | Duration |
|---|---|---|
| La Marzocco Manufacturer | 3 bar or line pressure | 2–8 s recommended |
| Slayer Manufacturer | adjustable from 0.5 bar, typically 2.5 bar | variable, part of the brew band |
| Decent Espresso Manufacturer | pressureless up to approx. 6.8 bar depending on profile | "Blooming" profile, 30 s and more |
| Lab setup Schmieder et al. 2023 Peer-reviewed | pre-infusion at 7 ml/s | main run 2.8–9.3 bar |
So the consensus range sits roughly at 0.5 to 4 bar and 2 to 10 seconds, with pressureless bloom variants as a special case. You can try out how pre-infusion, peak pressure and a declining profile affect extraction in the pressure profiling simulator.
The body myth, fact-checked
We searched the academic literature, manufacturer material and the community. No publication, no manufacturer claim, no measurement series reports a doubling of body through pre-infusion. Not even manufacturers with a commercial interest in it claim that. And the only direct measurement we know of on pre-infusion time and mouthfeel (10 versus 20 versus 30 seconds) found more sweetness but less mouthfeel with longer pre-infusion, plus visibly less crema. The sample was small, and we do not want to turn it into a counter-rule, but the direction simply does not hold.
There is a second point where two things often get blended together in the scene. Cameron et al. (2020) show very well what channeling costs: at the finest grind settings, up to 13.1 percent of the bed surface was no longer reached. But they do not show that pre-infusion prevents channeling. That bridge is plausible and every manufacturer claims it, but it has not been measured.
Community measurementWhat does exist is a careful measurement series across 24 shots with live pressure, flow and scale data. Its core finding: the usual pre-infusion steps are not enough to fully saturate the puck before extraction. Adding a ten-second pause after pre-infusion clearly lowered the peak resistance of the bed. That fits exactly with the 5 to 10 seconds from the X-ray measurement.
So here is what you can say with a clear conscience: pre-infusion makes sure the puck is saturated before full pressure arrives. More even saturation is plausible, and the physical basis for it has been measured cleanly. That the body doubles has not. That is our reading of the evidence, and we would be glad if someone knows a study we have overlooked.
Why fresh coffee behaves differently
The Coffee Excellence Center at ZHAW measured the degassing of roasted beans over time. The numbers explain a lot:
| Roast level | CO₂ released (whole beans) |
|---|---|
| light | 2.5–2.8 mg/g |
| medium | 4.7–6.6 mg/g |
| dark | 8.5–11.9 mg/g |
On top of that, a finding that hits practice directly: grinding releases up to 75 percent of the trapped gas within 80 seconds. And light roasts degas the most slowly, at 35 degrees degassing was still measurable a month after roasting. That is also why we do not state a roast date in days: the right resting time depends on the roast level, not on the calendar.
The link to pre-infusion is directly documented as far as the physics goes: the water enters a medium saturated with air and residual CO2, and has to displace that gas. The more residual gas, the more there is to displace. That this gas builds up counter-pressure with very fresh coffee, makes wetting uneven and thereby encourages channeling is the plausible continuation, but it has not been shown under controlled conditions. We label that a hypothesis, because that is what it is. You can play through how roast level and freshness work out in the roast level and freshness simulator.
Frequently asked questions
How long should pre-infusion last?
X-ray measurements show that the water front needs 5 to 10 seconds to work through the puck and displace the air. Manufacturer recommendations mostly sit at 2 to 8 seconds, and considerably longer for pressureless bloom profiles. A universally optimal duration is not documented, so we suggest testing with your own coffee.
Does pre-infusion double the body of an espresso?
No, there is no source for it. Neither the scientific literature nor manufacturer material contains such a measurement. The only direct measurement series we know of on pre-infusion time and mouthfeel actually found more sweetness but less mouthfeel and less crema with longer pre-infusion.
What happens inside the coffee puck during pre-infusion?
The water enters a pore network saturated with air and residual carbon dioxide and displaces that gas. At the same time the particles swell by around 15 percent. As a result the effective pore diameter drops from about 200 to 6 micrometres and the permeability by roughly a factor of a thousand.
Does pre-infusion prevent channeling?
That is plausible, but not measured. What is documented is what channeling costs: at fine grind settings, up to 13.1 percent of the bed surface was no longer reached. That pre-infusion prevents this is claimed by manufacturers, but so far it is not backed by a controlled study.
Does very freshly roasted coffee need a longer pre-infusion?
Very fresh coffee contains more residual CO₂, which has to be displaced during brewing. That a gentler or longer pre-infusion helps for that reason is common practice and physically plausible, but not documented by a controlled study. Interesting: light roasts contain less CO₂ in absolute terms, but degas the most slowly.
Roastery tip
An experiment that shows more than any theory: take the same coffee, once three days after roasting and once after two weeks, and run both with an identical profile. With Cozy Chocolate and Nutty Delight you see the difference in the flow straight away. Wild Peach shows it in the glass. Note down what you see, your own data beats any rule of thumb.
