CoffeariumBeta
  1. Plant
  2. Farm
  3. Process
  4. Trade
  5. Roast
  6. Brew
  7. Cup
Roasting · Science

Water in roasting

Where the water in a roast comes from and where it goes – the green coffee's own moisture, water formed by reactions, the heat needed to evaporate it, the steam that swells the bean, and the little that is left at the end.

Across the chain: Science

Source-checkedby Coffearium verification pass· Last source-checked 19 Oct 2026· Prepared with AI assistance· 16 sources (12 tier A)· 2 contested· Updated 19 Oct 2026
Reading depth
Contents
  • Green coffee enters the roaster holding about 10–12% water (wet basis in a textbook; basis not stated in a 2002 study).1,4
  • Reactions form more water. In a fluid-bed study, 10.2–11.4 g water per 100 g green coffee evaporated, more than the 8.3 g the beans brought in.2
  • Heating and evaporating this water takes much energy, so the early roast is mainly endothermic.1,5
  • Water vapour and carbon dioxide build the pressure that swells the bean; steam pressure is described as the cause of first crack.3,5
  • Little water stays in the bean; one study measured 2.56 g/100 g at light-medium and 0.89 g at very dark roast (basis not stated). Quenching can add more.6,13

Two sources of water#

Water enters the roaster with the green coffee. A roasting textbook chapter gives a typical 10–12% (wet basis, wb), and a 2002 roasting study gives the same range (basis not stated).1,4 By the end of the roast only a few percent is left.4 How this moisture is measured in green coffee is explained under moisture and water activity.

The roast also makes new water. Chemical reactions in the bean produce it, and it evaporates too.1,2 These reactions include Strecker degradation, the breakdown of chlorogenic acids and sucrose, and the Maillard reaction.2 Researchers call this reaction water, to set it apart from the green coffee’s own water.2

A 2005 Swiss laboratory study measured both. It roasted 100 g batches with 8.3 g water in a fluid bed, once fast and once slow.2 More water left the roaster than the beans had brought in: 10.2 g in the fast roast and 11.4 g in the slow one.2 Reaction water made up 41% and 36% of it.2

024681012g water per 100 g green coffeeBrought in by the greencoffee(8.3 g/100 g, wb)8.3 g2 Evaporated in the roaster,fast roast(170 s, 260 °C air)10.2 g2 Evaporated in the roaster,slow roast(720 s, 228 °C air)11.4 g2 Of which reaction water,fast roast(41%)≈4.2 g (our calculation)2 Of which reaction water,slow roast(36%)≈4.1 g (our calculation)2
Data table
g water per 100 g green coffeeSource
Brought in by the green coffee (8.3 g/100 g, wb)8.3 g2
Evaporated in the roaster, fast roast (170 s, 260 °C air)10.2 g2
Evaporated in the roaster, slow roast (720 s, 228 °C air)11.4 g2
Of which reaction water, fast roast (41%)≈4.2 g (our calculation)2
Of which reaction water, slow roast (36%)≈4.1 g (our calculation)2
Water per 100 g green coffee in a laboratory fluid bed: what the beans brought in, what evaporated in the roaster in a fast (170 s, 260 °C air) and a slow (720 s, 228 °C air) roast to the same colour, and the share formed by reactions. Reaction water in grams is our calculation from the reported totals and percentages.2

How water moves through the roast#

Heating and evaporation. At first, much of the heat goes into warming the water in the bean. The early roast is therefore mainly endothermic: it takes in a lot of energy.5 Evaporation then draws more heat. A 2025 single-bean model places the start of this evaporative cooling at 100 °C.8

Inside the bean. Water turns to vapour inside the bean, not only at its surface. One 2017 model assumes evaporation inside the bean and predicts a drying front moving towards the centre.9 The first stage of the roast, up to the visible colour change, is covered under drying phase.

Water keeps leaving. In the 2005 study, the green coffee’s own water left first.2 Reaction water and carbon dioxide rose clearly only once the bean core reached 180–200 °C.2 In the slow roast, the original water had almost stopped leaving after about 300 s of 720 s; reaction water then dominated.2

Pressure and first crack. Vapour cannot escape the dense bean quickly, so pressure builds inside it.3 Together with carbon dioxide, this pressure makes the bean expand1,3; an SCA handbook adds that it starts the bean cracking.3 A 2020 sensory paper describes first crack as accumulated steam pressure breaking the beans and expelling steam.5

1Green coffee inAbout 10–12% water(wb in a textbook) 1,4 2HeatingWarming the water takesmuch energy 5 ; evaporativecooling from 100 °C 8 3Evaporation insideDrying front movesto the centre(one model) 9 4Reaction waterFormed from a bean coreof 180–200 °C; 36–41%of water evaporated 2 5Pressure and swellingVapour and CO₂ swellthe bean; steam pressureat first crack 1,3,5 6DropAbout 1–2.5% left 5 ;quenching may addwater 13 water vapour outwater vapour outwater vapour outheat incoffee through the roastheat in, vapour outSchematic, not to scale. Values from single studies under their own conditions.
The path of water through a roast, from the green coffee to the drop, with water vapour leaving along the way. Numbers are from single studies under their own conditions.1,2,3,4,5,8,9,13 Schematic, not to scale.

What water does in the bean#

It carries the reactions. According to a roasting textbook chapter, the speed of key flavour reactions depends on how much water is available.1 Some reactions slow down once the moisture falls below a critical value.1 A chemistry chapter in the same book describes the roasting bean as a low-moisture reaction system.11

It softens the bean. The cell walls can be rigid (“glassy”) or soft (“rubbery”), depending on moisture and temperature.1 According to the textbook chapter, the bean swells while rubbery, so the speed of drying helps decide how far it expands.1

It sets the profile’s options. Two roasts can reach the same colour in the same time by different paths.1 According to the chapter, one path may evaporate more water early than the other. This creates different combinations of bean temperature and water activity, which change the reactions.1

Green coffee moisture and the roast#

Wetter green coffee loses more water in the first phase of the roast. According to the textbook chapter, it ends at a similar final moisture, so its roast loss is higher.1 Roast loss then varies between batches roasted to the same colour.1

The effect on flavour is described as strongest in light roasts. In dark roasts, most aroma differences from initial moisture level off.11 The authors of a 2025 roasting model recommend a low initial moisture to make the roast easier to control.8

Water can also be added deliberately. According to the chemistry chapter, steam-treated green coffee reaches its target colour faster than untreated coffee.11 In a 2019 study, robusta roasted in superheated steam heated up, lost moisture and darkened faster than in hot air.12

Water left at the end#

Roasted coffee is dry. A 2020 multi-study paper gives 1–2.5% moisture for roasted beans (basis not stated).5 Darker roasts hold less. In a 2021 study of one Ethiopian arabica, moisture fell from 2.56 g/100 g at light-medium to 0.89 g/100 g at very dark roast (basis not stated).6

Roast speed matters too. According to the textbook chapter, fast roasts end with slightly more water than slow roasts to the same colour.1 A doctoral thesis from a Swiss laboratory reported the same direction.7

012345678910g water per 100 g (basis not stated in Kim et al. and Münchow et al.)Green coffee(Ethiopian arabica)8.646 Light-medium roast2.566 Medium roast2.176 Moderately dark roast1.296 Very dark roast0.896 Roasted beans, literaturerange1–2.55 Single beans, 200–280 °Cfor 10 min(model study)2–48
Data table
Pointg water per 100 g (basis not stated in Kim et al. and Münchow et al.)Source
Green coffee (Ethiopian arabica)8.648.646
Light-medium roast2.562.566
Medium roast2.172.176
Moderately dark roast1.291.296
Very dark roast0.890.896
Roasted beans, literature range1–2.51–2.55
Single beans, 200–280 °C for 10 min (model study)2–42–48
Water left in the beans: one Ethiopian arabica from green to very dark roast, oven-dried at 105 °C6; a literature range for roasted beans5; and single beans in a model study.8 Basis not stated in the first two sources.

Water quenching can add water back after the roast. In industrial trials, different amounts of quench water gave coffees with 2.3–8.8 g water per 100 g (wb).13 What this changes is explained under cooling and quenching.

Reference values#

Values reported in the sources. They describe particular coffees, roasters and models; they are not targets.

Parameter Value reported Source
Green coffee moisture, typical 10–12% (wb); 10–12% (basis not stated) Schenker & Rothgeb1, Yeretzian et al.4
Green coffee moisture in the studies cited 8.3 g/100 g (wb); 8.64 g/100 g (basis not stated); 0.1296 kg/kg dry basis, ≈11.5% wb (our conversion) Geiger et al.2, Kim et al.6, Bustos-Vanegas et al.8
Green coffee adjusted for roasting trials 3.2–18.2 g/100 g (wb), from an original 11.1 g/100 g Schenker7
Water evaporated in the roaster, fluid bed, per 100 g green coffee (8.3 g in) 10.2 ± 1.2 g (fast, 170 s, 260 °C air); 11.4 ± 1.4 g (slow, 720 s, 228 °C air) Geiger et al.2
Share formed by reactions 41% (fast); 36% (slow); ≈4.2 g and ≈4.1 g (our calculation) Geiger et al.2
Water leaving during cooling, estimated 1.6 g (fast); 0.1 g (slow) per 100 g green coffee Geiger et al.2
Carbon dioxide leaving in the roaster, same study 0.4 g (fast); 0.50 g (slow) per 100 g green coffee Geiger et al.2
Reaction water and CO₂ start Bean core 180–200 °C Geiger et al.2
Green coffee’s own water almost gone After about 300 s of a 720 s roast Geiger et al.2
Evaporative cooling, single-bean model Starts at 100 °C, decreasing to the end of the roast Bustos-Vanegas et al.8
Latent heat of vaporisation used in one model 2.3 × 10⁶ J/kg Fadai et al.9
Moisture of roasted beans, general 1–2.5% (basis not stated); about 2.5% (basis not stated) Münchow et al.5, Schenker & Rothgeb1
Moisture by roast degree, one Ethiopian arabica (oven, 105 °C) Green 8.64; light-medium 2.56; medium 2.17; moderately dark 1.29; very dark 0.89 g/100 g (basis not stated) Kim et al.6
Final moisture, single beans (gas-fired drum, 200–280 °C for 10 min; model study) 2–4%, lower with longer time and higher temperature Bustos-Vanegas et al.8
Water after quenching, industrial trials 2.3–8.8 g/100 g (wb) Baggenstoss et al.13
Glass transition, calorimetry, aw 0.155 → 0.512 Green: 48.76 → 34.89 °C; roasted: 45.73 → 40.15 °C Iaccheri et al.10
Superheated steam roasting, robusta 190–250 °C, compared with hot air at the same temperatures Chindapan et al.12
Moisture method, green coffee Loss in mass at 105 °C; about 1.0% lower than the ISO 1446 reference method (2003 edition, since replaced) ISO 6673:200314, ISO 6673:202516
Moisture method, roasted ground coffee Loss in mass at 103 °C (routine method); best suited to degassed coffee ISO 1129415

Physics: heat, vapour and pressure#

Energy. Heating water and turning it into vapour is the main energy demand of the early roast.1,5 One single-bean model uses a latent heat of vaporisation of 2.3 × 10⁶ J/kg.9 In a 2025 lumped model, evaporation affected the modelled bean temperature more than the heat released by reactions.8 A 2002 study describes the bean temperature rising again towards 170 °C once the beans reach about 6% moisture.4 Where the roast turns exothermic is covered under endothermic and exothermic phases.

Where water evaporates. The 2017 model assumes that liquid water evaporates only inside the bean, at the interface between liquid and gas.9 The authors judge this local evaporation more realistic than earlier approaches, but name the lack of physical data as the model’s main drawback.9 The model tracks carbon dioxide but has no term for water formed by reactions (our reading of its equations).9 Only the gas measurements of the 2005 study separate the two kinds of water.2

Mass transfer. According to the roasting textbook chapter, a higher air-to-bean ratio raises air speed at the surface and speeds up dehydration.1 The chapter adds that water takes time to redistribute within the bean, which may limit dehydration in short roasts.1 The chapter gives this next to the slightly higher final moisture of fast roasts, which a doctoral thesis also reported.1,7

Online evidence. A 2002 study followed the roaster headspace by mass spectrometry, roasting 40 g of Colombian arabica isothermally at 180–190 °C.4 The water signal rose first, reflecting the drying of the beans, and then fell.4 In the 2005 fluid-bed study, the peak evaporation rate was higher in the fast roast, while more water evaporated in total in the slow one.2

Pressure. According to an SCA handbook, vapour builds internal pressure in the dense green coffee and is only released once the temperature is high enough.3 According to the textbook chapter, the gas together with the water vapour drives the expansion of the bean.1 No study opened for this entry measured the vapour pressure inside a roasting bean.

Chemistry: water made and water used#

Where reaction water comes from. The 2005 study names Strecker degradation, degradation of chlorogenic acid and sucrose, decarboxylation of amino acids and the Maillard reaction as the sources of carbon dioxide and reaction water.2 According to the chemistry chapter, roasting first drives off free water and then enters a dry heating stage with dehydration, fragmentation and pyrolysis reactions.11

How it was measured. The study roasted beans pre-dried to 1.1 g/100 g (wb) next to normal beans at 8.3 g/100 g.2 Water from the pre-dried beans was taken as reaction water; the difference gave the green coffee’s own water.2 Late in the roast, reaction water evaporation declined. The authors suggest the substrate ran out, or the water was used up in other reactions inside the cells.2

Water activity and reactions. Temperature, water activity and pressure govern the kinetics of the flavour reactions inside the intact bean, according to the textbook chapter.1 A different early profile can therefore change the flavour at the same colour and time.1 Initial moisture is described as mattering mainly in light roasts.11

Steam before or during roasting. According to the chemistry chapter, steam treatment of green robusta changes sugars and amino acids, reduces some undesired compounds and shortens the time to the target colour.11 In the 2019 superheated steam study, steam-roasted robusta had a lower pH and more sucrose, glucose and arabinose than hot-air roasted beans.12 A doctoral thesis built steam injection into its laboratory roaster for humid-atmosphere roasting.7 Its results on humid air were not in the part of the thesis opened for this entry.

Glass transition and moisture#

Water lowers the glass transition temperature of coffee. In calorimetry on samples held at set water activities, it fell from 48.76 °C to 34.89 °C in green coffee and from 45.73 °C to 40.15 °C in roasted coffee as water activity rose from 0.155 to 0.512.10 Both green and roasted beans are mainly glassy.10

During roasting, the evidence is descriptive. The textbook chapter describes a passage from glassy to rubbery and back, governed by moisture and temperature.1 An SCA handbook places the transition around first crack, with cells inflated by carbon dioxide.3 No measurement inside a roasting bean was found (see Contested).

Measuring water before and after#

Green coffee. ISO 6673 determines loss in mass at 105 °C. By convention it serves as a water content method, giving results about 1.0% lower than the ISO 1446 reference method.14 These details are from the 2003 edition. ISO has since withdrawn it and published a third edition in 2025, whose full text was not opened for this entry.16

Roasted coffee. ISO 11294 is a routine method for roasted ground coffee, by loss in mass at 103 °C.15 ISO describes it as most suited to degassed coffee, because carbon dioxide and other volatiles also leave on heating.15 Research methods vary: the 2005 study dried finely ground roasted coffee at 103 °C for 5 h, and the 2021 study oven-dried at 105 °C.2,6

During the roast. The 2005 study measured water vapour in the exhaust air with an infrared gas analyser.2 The authors expect the method could be adapted to larger roasters.2

Risks and defects#

  • Varying green moisture. It makes the roast harder to control and roast loss a weaker guide to roast degree.1,8
  • Steep gradients in very fast roasts. According to the textbook chapter, the surface may over-roast while the core stays under-roasted.1
  • Over-quenching. Excess quench water is absorbed and raises the final moisture.1,13
  • Misleading readings. Oven methods on fresh roasted coffee also count escaping volatiles; ISO 11294 is best suited to degassed coffee.15

Open questions#

  • Moisture inside the bean. Models place evaporation inside the bean, but their authors lack physical data to fit them.9
  • Reaction water in production roasters. The only split into own and reaction water opened for this entry comes from a 100 g laboratory fluid bed.2
  • Initial moisture and flavour. The effect is described in a textbook chapter11; the underlying study was not opened for this entry.
  • Humid-air roasting. Superheated steam was tested on robusta at laboratory scale12; no sensory study of humid-air roasting was found.
  • Vapour pressure at first crack. It is described but was not measured in any source opened.3,5

Key studies#

Geiger et al., 2005 – Water and carbon dioxide in the exhaust. Measures water vapour and carbon dioxide leaving a laboratory fluid bed during a fast and a slow roast to the same colour, and roasts pre-dried beans to isolate reaction water. More water evaporated than the beans brought in; reaction water made up 36–41% of it and appeared from a bean core temperature of 180–200 °C.

Journal of Food Science 70(2), E124–E130 · doi:10.1111/j.1365-2621.2005.tb07084.x

Yeretzian et al., 2002 – Online monitoring of roasting. Follows volatiles and water in the headspace of a small isothermal roaster by proton-transfer-reaction mass spectrometry. The water signal rose first as the beans dried and then fell, while other compounds appeared later.

European Food Research and Technology 214, 92–104 · doi:10.1007/s00217-001-0424-7

Kim et al., 2021 – An Ethiopian arabica at four roast degrees. Roasts one coffee from light-medium to very dark and measures moisture, colour, weight loss and extract chemistry. Moisture fell with each step of roast degree, from 8.64 g/100 g green to 0.89 g/100 g at very dark.

Food Science and Biotechnology 30(2), 235–244 · doi:10.1007/s10068-020-00865-w

Chindapan et al., 2019 – Superheated steam versus hot air. Roasts robusta at 190–250 °C in superheated steam and in hot air. Steam speeded up heating, moisture loss and darkening, and gave beans with a lower pH and more of some sugars.

Journal of Food Science 84(2), 292–302 · doi:10.1111/1750-3841.14422

Bustos-Vanegas et al., 2025 – Evaporation and heat generation. Models the temperature of single beans in a drum roaster, with and without evaporation and reaction heat. Evaporation had the greater effect on the modelled temperature, and the authors recommend a low initial moisture for easier control.

Frontiers in Food Science and Technology 5, 1603783 · doi:10.3389/frfst.2025.1603783

Contested
  • The main gas of roasting: an SCA handbook names water vapour as the main component of the gas produced inside the bean3; a roasting textbook chapter names carbon dioxide as the predominant gas formed.1 The two count different things: the chapter speaks of gas from pyrolysis and the Maillard reaction, while by mass a fluid-bed study measured far more water (10.2–11.4 g) than carbon dioxide (0.4–0.50 g) leaving per 100 g green coffee in the roaster.2
  • When the bean softens: a roasting textbook chapter describes the cell walls passing from glassy to rubbery and back to glassy as moisture and temperature change1; an SCA handbook places the glass transition around first crack.3 The calorimetry study opened for this entry measured it in green and roasted coffee held at set water activities, not during a roast.10
Share or cite this entry

Change log

  • – First crack attributed to the SCA handbook alone; the roasting textbook chapter supports only the expansion of the bean. Reason: Verification pass; the chapter describes cracking as a microburst of gas, not as caused by vapour pressure.
  • – ISO 6673 details marked as from the withdrawn 2003 edition, with the 2025 edition added as a source. Reason: Verification pass; ISO lists ISO 6673:2003 as withdrawn and replaced by ISO 6673:2025.

Sources

  1. BSchenker, S.; Rothgeb, T. (2017). The Roast – Creating the Beans' Signature. In: Folmer, B. (ed.), The Craft and Science of Coffee, Academic Press, ch. 11, pp. 245–271. doi:10.1016/B978-0-12-803520-7.00011-6 · other copies
  2. AGeiger, R.; Perren, R.; Kuenzli, R.; Escher, F. (2005). Carbon Dioxide Evolution and Moisture Evaporation During Roasting of Coffee Beans. Journal of Food Science 70(2), E124–E130. doi:10.1111/j.1365-2621.2005.tb07084.x · other copies
  3. BSmrke, S.; Sage, E.; Wellinger, M.; Yeretzian, C. (2018). The Coffee Freshness Handbook (ISBN 978-0-9995807-2-1). Specialty Coffee Association, Santa Ana.
  4. AYeretzian, C.; Jordan, A.; Badoud, R.; Lindinger, W. (2002). From the green bean to the cup of coffee: investigating coffee roasting by on-line monitoring of volatiles. European Food Research and Technology 214, 92–104. doi:10.1007/s00217-001-0424-7 · other copies
  5. AMünchow, M.; Alstrup, J.; Steen, I.; Giacalone, D. (2020). Roasting Conditions and Coffee Flavor: A Multi-Study Empirical Investigation. Beverages 6(2), 29. doi:10.3390/beverages6020029 · other copies
  6. AKim, I.; Jung, S.; Kim, E.; Lee, J.-W.; Kim, C.-Y.; Ha, J.-H.; Jeong, Y. (2021). Physicochemical characteristics of Ethiopian Coffea arabica cv. Heirloom coffee extracts with various roasting conditions. Food Science and Biotechnology 30(2), 235–244. doi:10.1007/s10068-020-00865-w · other copies
  7. BSchenker, S. (2000). Investigations on the Hot Air Roasting of Coffee Beans (Diss. ETH No. 13620). ETH Zürich (doctoral thesis).
  8. ABustos-Vanegas, J. D.; Martins, M. A.; Corrêa, P. C.; Baptestini, F. M.; de Oliveira, G. H. H. (2025). Modeling and simulation of coffee bean heating during roasting: effect of heat generation. Frontiers in Food Science and Technology 5, 1603783. doi:10.3389/frfst.2025.1603783 · other copies
  9. AFadai, N. T.; Melrose, J.; Please, C. P.; Schulman, A.; Van Gorder, R. A. (2017). A heat and mass transfer study of coffee bean roasting. International Journal of Heat and Mass Transfer 104, 787–799. doi:10.1016/j.ijheatmasstransfer.2016.08.083 · other copies
  10. AIaccheri, E.; Ragni, L.; Cevoli, C.; Romani, S.; Dalla Rosa, M.; Rocculi, P. (2019). Glass transition of green and roasted coffee investigated by calorimetric and dielectric techniques. Food Chemistry 301, 125187. doi:10.1016/j.foodchem.2019.125187 · other copies
  11. BPoisson, L.; Blank, I.; Dunkel, A.; Hofmann, T. (2017). The Chemistry of Roasting – Decoding Flavor Formation. In: Folmer, B. (ed.), The Craft and Science of Coffee, Academic Press, ch. 12, pp. 273–309. doi:10.1016/B978-0-12-803520-7.00012-8 · other copies
  12. AChindapan, N.; Soydok, S.; Devahastin, S. (2019). Roasting Kinetics and Chemical Composition Changes of Robusta Coffee Beans During Hot Air and Superheated Steam Roasting. Journal of Food Science 84(2), 292–302. doi:10.1111/1750-3841.14422 · other copies
  13. ABaggenstoss, J.; Perren, R.; Escher, F. (2008). Water content of roasted coffee: impact on grinding behaviour, extraction, and aroma retention. European Food Research and Technology 227(5), 1357–1365. doi:10.1007/s00217-008-0852-8 · other copies
  14. AInternational Organization for Standardization (2003). ISO 6673:2003 Green coffee — Determination of loss in mass at 105 °C. ISO.
  15. AInternational Organization for Standardization (1994). ISO 11294:1994 Roasted ground coffee — Determination of moisture content — Method by determination of loss in mass at 103 °C (Routine method). ISO.
  16. AInternational Organization for Standardization (2025). ISO 6673:2025 Green coffee — Determination of loss in mass at 105 °C. ISO.

Tier A · authoritative · B · trade and expert · C · general. Supporting quotes for every source are in the data version of this entry. If a publisher’s site refuses access, try the doi link or “other copies”.

Suggest a change

On Water in roasting. Tell us what you would change and why. A source that supports it helps most.