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

Glass–rubber transition

How the material of the coffee bean switches between a rigid, glassy state and a soft, rubbery one, how water and heat drive the change, and what it means for swelling in the roaster, grinding and storage.

Across the chain: Science

Source-checkedby Coffearium verification pass· Last source-checked 20 Oct 2026· Prepared with AI assistance· 15 sources (10 tier A)· 3 contested· Updated 20 Oct 2026
Reading depth
Contents
  • A glass transition is the change of an amorphous material from a brittle, glassy state to a soft, rubbery one.3,5
  • Water lowers the glass transition temperature (Tg). In one study, Tg of green coffee fell from 48.76 °C to 34.89 °C as water activity rose from 0.155 to 0.512.2
  • According to a roasting textbook chapter, the bean passes from glassy to rubbery and back to glassy during roasting, and swells while rubbery.1
  • Roasted beans are brittle; added water makes them firmer and less brittle.10,11,12
  • No study opened for this entry measured the transition inside a roasting bean; published Tg values for coffee also disagree widely.2,4

What a glass transition is#

Many dry foods are amorphous: their molecules are packed without the order of a crystal. Such a material can be in one of two states.3 Below a certain temperature it is glassy, rigid and brittle. Above it, the material turns rubbery and soft.3,5

The change happens over a temperature range, the glass transition temperature (Tg).3 It is reversible: cooling takes the material back to the glassy state.3

Water moves Tg down. Food scientists call water a plasticiser: it softens the material, much as heat does.3 In dry foods, a little extra water lowers Tg sharply.3 So the state of a dry food depends on two things at once: temperature and water content.

Glass and rubber in the coffee bean#

In the coffee bean, the material that matters is the cell wall. The walls are thick and made of polysaccharides.1 According to a roasting textbook chapter, these polysaccharides are glassy or rubbery depending on moisture and temperature.1 The chapter adds that the change is gradual and blurred, because the wall holds several polysaccharides with different transitions.1

Stored green and roasted coffee beans are both mainly glassy.2 In a 2019 calorimetry study, Tg of green coffee fell from 48.76 °C to 34.89 °C as water activity rose from 0.155 to 0.512.2 Roasted coffee was less sensitive: its Tg fell from 45.73 °C to 40.15 °C over the same range.2

Not every study agrees. A 2021 study of ground Brazilian arabica found Tg midpoints near −11 °C in green, light and dark roasted samples.4 Its authors also describe water as a plasticiser of coffee, but their three samples differed by less than 1 °C.4 The two studies disagree on the level by about 50 °C at similar water activity (see Contested).2,4

0.10.20.30.40.50.6-30-20-100102030405060Water activity (aw)Glass transition temperature Tg (°C)Roasted, DSC (2019)2 Green, DSC (2019)2 Green, MDSC (2021)4 Dark roast, MDSC (2021)4 Light roast, MDSC (2021)4
Data table
SeriesWater activity (aw)Glass transition temperature Tg (°C)Source
Green, DSC (2019)0.15548.762
Green, DSC (2019)0.51234.892
Roasted, DSC (2019)0.15545.732
Roasted, DSC (2019)0.51240.152
Green, MDSC (2021)0.486-11.244
Light roast, MDSC (2021)0.399-11.934
Dark roast, MDSC (2021)0.388-11.874
Glass transition temperatures of coffee reported against water activity: green and roasted coffee by DSC in a 2019 study (end points of the range only)2, and green, light and dark roasted Brazilian arabica by modulated DSC in a 2021 study (ground samples, midpoints of the transition).4 The 2019 abstract does not say which point of the transition it reports. Lines only join the two reported points; the gap between the studies is unexplained.

What happens in the roaster#

According to the textbook chapter, the bean passes from glassy to rubbery and finally back to glassy during roasting.1 Heat pushes the bean towards the rubbery state. Drying pulls the other way, because a drier material has a higher Tg.1,3 According to the chapter, the speed of drying therefore plays a key role.1

The bean swells while it is rubbery. Gas pressure inside the bean drives the expansion, and the thick cell walls resist it.1 In the rubbery state the walls resist less, so they stretch.1 An SCA handbook describes the same picture: the material softens and carbon dioxide inflates the cells.8 Where these gases come from is covered under roast gases and CO₂.

The walls stretch but do not burst. According to the textbook chapter, the framework of cell walls stays intact while the cells empty and fill with gas.1 A doctoral thesis links the absence of cell-wall rupture, probably, to the rubbery state of the walls at high temperature.14 The result is the open, porous bean described under porosity and cell structure.

1Green coffee: glassyDense tissue, thickrigid walls 1 Water lowers Tg 2,3 2Roasting: rubberyWalls soften and stretch 1 Gas pressure swellsthe cells 1,8 Walls stay intact 1,14 3Roasted, cooled: glassyPorous and brittle 1,11 Crushed, then ground 14 Added water: firmer,less brittle 10,12 heatcoolSchematic, not to scale. When each change happens in a roast has not been measured.
Cell walls of the bean in three states: rigid and glassy in green coffee, soft and stretched by gas pressure during roasting, and glassy, porous and brittle after cooling.1,2,3,8,10,11,12,14 Schematic, not to scale.

Why roast speed changes the bean#

Because the state depends on both temperature and water, the speed of drying matters. According to the textbook chapter, the roast profile is therefore crucial for how far the bean swells.1

A 2000 study compared a fast, hot roast with a slow, cooler one to the same roast degree. The fast roast gave beans with greater volume, more pore space and larger micropores in the cell walls.6 A doctoral thesis from the same laboratory reviews a patent that explains this with water. Fast roasts retain more water at high temperature, and water softens the bean.7 This explanation was not tested directly in any source opened for this entry.

How the swelling is timed against first crack is still open. The textbook chapter and the SCA handbook place the transition differently (see Contested), and first crack has its own entry.1,8

After the roast: brittle beans#

Roasted beans are brittle. In a 2001 texture study, roasting made beans lose strength and toughness and become progressively more brittle.11 The authors relate this to water loss and to the changed tissue structure.11 According to a doctoral thesis, their brittleness means roasted beans are first crushed into fragments and then ground finer.14

Water reverses part of this. Beans with more water after quenching were firmer in a shear test.12 In a 2007 texture study, plasticisation by water took away the typical brittleness of roasted beans.10 Wetter roasted coffee also ground coarser and needed a rest before grinding.13

This links the transition to quench water. A small amount of quench water mainly cools the bean; too much is absorbed and raises its water content.1,13 How quenching is done is explained under cooling and quenching.

Reference values#

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

Parameter Value reported Source
Tg, green coffee, DSC 48.76 °C at aw 0.155; 34.89 °C at aw 0.512 Iaccheri et al.2
Tg, roasted coffee, DSC 45.73 °C at aw 0.155; 40.15 °C at aw 0.512 Iaccheri et al.2
Tg midpoint, ground Brazilian arabica, modulated DSC Green −11.24 °C (aw 0.486); light roast −11.93 °C (aw 0.399); dark roast −11.87 °C (aw 0.388); endpoints −1.69 to −2.49 °C Brzezińska et al.4
Tg of pure water, used in food state diagrams −135 °C Sablani et al.3
Tg of dry bean material, model parameter ≈190 °C Fadai et al.5
Water activity where hardening by water ends · where further softening begins (compression tests) Green 0.61 · 0.78; roasted 0.75 · 0.86 Rocculi et al.9
Critical water activity above which beans soften Raw 0.75; roasted about 0.86 Pittia et al.10
Firmness of quenched beans, light roasts (industrial, L* 29–30), 2.4–5.3 g/100 g water (wb) Fmax = 2777 + 866 × water content (N; r² = 0.998) Baggenstoss et al.12,14
Firmness of quenched beans, dark roasts (laboratory, L* 21–23), 1.7–4.2 g/100 g water (wb) Fmax = 3575 + 231 × water content (N; r² = 0.949) Baggenstoss et al.12,14
Water content after quenching, industrial trials 2.3–8.8 g/100 g (wb) Baggenstoss et al.13

Physics: plasticisation and the state diagram#

Plasticisation. Water has a low molecular weight and a low Tg of −135 °C, so it depresses the Tg of food solids.3 It adds free volume between the large molecules and lets them relax.3 Heat has a similar effect, by raising molecular energy.3

Gordon–Taylor relation. For a binary system of food solids and water, a 2010 review gives the Gordon–Taylor equation.3 It estimates Tg of the mixture from the Tg of the dry solids and of water, their proportions and a fitted constant.3 No Gordon–Taylor fit for coffee bean material was found for this entry.

State diagram. A state diagram plots the physical state of a food against temperature and water content.3 Below the Tg curve the material is glassy; above it, rubbery.3 According to the roasting textbook chapter, coffee has several such lines, one for each cell-wall polysaccharide.1 In the chapter’s account, a roast is a path across this diagram: temperature rises while water content falls.1 No study opened for this entry drew that path on measured Tg values of coffee.

A model of the roasting bean. A 2019 model of stresses in a roasting bean uses a Tg that falls with water content, from about 190 °C for dry material.5 Below Tg the model treats the wall material as elastic; above it, as viscous and permanently deformable.5 In the simulations, the wetter interior passes through the transition first and becomes mechanically weak.5 A large spike in stress follows, which the authors suggest may signal the onset of first crack.5 The dry-material value is a model parameter, for which the authors point to a conference paper; it was not measured in that study.5

Water: plasticiser and antiplasticiser#

At low water contents, water does not always soften coffee. A 2011 study combined compression tests, calorimetry and NMR on green and roasted beans.9 Up to a water activity of 0.61 (green) and 0.75 (roasted), water hydrated the solid matrix and hardened the beans.9 Only once this hydration was complete did water act as a plasticiser and the beans soften.9 Above 0.78 (green) and 0.86 (roasted), hardness fell and the elastic modulus fell further.9

A 2007 study on the same kind of beans found the matching pattern. Strength and toughness rose with hydration up to a critical water activity, 0.75 for raw and about 0.86 for roasted beans; above it, the bean matrix progressively softened.10

Quenched coffee fits the hardening side. In industrial and laboratory trials, the maximum shear force of roasted beans rose almost linearly with water content.12,14 The rise was steeper in light roasts than in dark ones, and light roasts were less brittle overall.14 The thesis calls this a plasticising effect of water.14 In the 2011 study’s terms, firmer beans at low water activity are antiplasticisation (see Contested).9

123456300040005000600070008000Water content of roasted beans (g/100 g, wb)Maximum shear force (N)Light roasts, industrial12 Dark roasts, laboratory12
Data table
SeriesWater content of roasted beans (g/100 g, wb)Maximum shear force (N)Source
Light roasts, industrial2.4485512
Light roasts, industrial5.3736712
Dark roasts, laboratory1.7396812
Dark roasts, laboratory4.2454512
Firmness of roasted beans (maximum force in a Kramer shear cell) against water content after quenching: light roasts from a 320 kg industrial roaster and dark roasts from a 200 g laboratory fluid bed. Lines are drawn from the reported regressions over the water contents produced in each set of trials; end values are our calculation.12,14

Grinding and storage#

Grinding. Wetter roasted coffee ground coarser, and it needed time to equilibrate before grinding gave good results.13 Drier coffee did not show this time dependence.13 In the thesis, the coarser grind is explained by the lower brittleness of wetter beans, which need more energy to break.14 According to the grinding chapter, beans with high moisture remain rather elastic and hard to grind even after a long rest.15 The same chapter notes that liquid nitrogen makes roasted beans more brittle.15

Storage. Green and roasted coffee are mainly glassy, and this glassy structure slowly changes during storage.2 In industrial trials, higher water content lowered the aroma stability of roasted coffee in closed storage.13 The thesis explains this most probably by plasticisation: water lowers Tg and lets reactants move faster.14 The thesis names this as the likely cause; the reaction mechanisms themselves were not shown.14 The authors of the 2021 study conclude from their Tg range that roasted beans can be kept at room temperature.4 Packaging, oxygen and staling are covered under freshness and storage.

Measuring the transition#

Calorimetry. In differential scanning calorimetry (DSC), the glass transition shows as a step in the heat-flow curve.3 Studies report its onset, midpoint or endpoint, and the value depends on the scan rate.3 The 2021 study used the modulated form (MDSC), which it describes as suited to glass transitions.4 The 2021 study reports onset, midpoint and endpoint of the step.4 The 2019 study added dielectric spectroscopy at 1.6–2.7 GHz and found it correlated with water activity and Tg.2

Mechanical tests. Mechanical methods follow the softening directly. A doctoral thesis on hot-air roasting used dynamic mechanical thermal analysis (DMTA) on green-coffee slices heated steadily from ambient temperature.7 Its results were not in the part of the thesis opened for this entry. Compression and shear tests give the hardness, modulus and firmness values above.9,10,12

Limits. All the measured Tg values come from stored beans at room-temperature water activities.2,4 During roasting the bean dries and changes chemically at the same time. No source opened for this entry measured Tg under those conditions.

Risks and defects#

  • Over-quenching. Excess water is absorbed and raises the water content.1,13 Wetter beans are firmer, grind coarser and need time to equilibrate before grinding.12,13
  • Shorter aroma stability. Higher water content lowered aroma stability in closed storage.13
  • Roast profile and swelling. According to the textbook chapter, swelling happens in the rubbery state, so the drying rate shapes bean volume.1 In one study, a fast, hot roast gave larger and more porous beans than a slow one.6

Open questions#

  • Transition during a roast. No measurement of Tg or of the state of the cell walls inside a roasting bean was found.1,2
  • Level of Tg in coffee. Two calorimetry studies differ by about 50 °C at similar water activity.2,4
  • Water retention and expansion. That fast roasts swell more because they stay wetter is an explanation, not a tested result, in the sources opened.6,7
  • Stress and first crack. A model links a stress spike at the transition to first crack; it has not been checked against measurements.5
  • Glass transition and staling. Plasticisation is the proposed reason for faster aroma loss in wetter coffee; the mechanisms are not known in detail.14

Key studies#

Iaccheri et al., 2019 – Glass transition of green and roasted coffee. Measures Tg of green and roasted beans held at set water activities by calorimetry, and relates dielectric spectra to it. Both are mainly glassy; water lowers Tg, more strongly in green than in roasted coffee.

Food Chemistry 301, 125187 · doi:10.1016/j.foodchem.2019.125187

Rocculi et al., 2011 – Water that hardens before it softens. Combines texture analysis, sorption isotherms, calorimetry and NMR on green and roasted beans. Water first hardens the matrix as it hydrates it, and acts as a plasticiser only above a critical water activity.

Journal of Agricultural and Food Chemistry 59, 8265–8271 · doi:10.1021/jf201333a

Schenker et al., 2000 – Pore structure and roasting conditions. Roasts beans in a fast, hot and a slow, cooler fluid-bed process to the same roast degree and studies volume, pores and microstructure. The fast roast gives larger beans, more pore volume and larger micropores in the cell walls.

Journal of Food Science 65(3), 452–457 · doi:10.1111/j.1365-2621.2000.tb16026.x

Baggenstoss et al., 2007 – Quench water and the roasted bean. Compares air-cooled and water-quenched coffee at laboratory and industrial scale. Wetter beans degassed faster and were firmer; most aroma compounds were similarly stable.

Journal of Agricultural and Food Chemistry 55(16), 6685–6691 · doi:10.1021/jf070338d

Fadai et al., 2019 – Stresses in a roasting bean. Models the bean as a material that turns from elastic to viscous at a moisture-dependent Tg. The wetter interior softens first, and a stress spike follows that the authors link to first crack.

International Journal of Non-Linear Mechanics 110, 123–130 · doi:10.1016/j.ijnonlinmec.2018.12.006

Contested
  • 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.8 Neither is a measurement inside a roasting bean.
  • Glass transition temperature of coffee beans: a 2019 study by differential scanning calorimetry found 34.89–48.76 °C in green and 40.15–45.73 °C in roasted coffee2; a 2021 study by modulated calorimetry found midpoints of −11.2 to −11.9 °C in ground green, light and dark roasted arabica.4 Even its endpoints, −1.7 to −2.5 °C, lie far below the 2019 values; the 2019 abstract does not say which point of the transition it reports or whether the beans were ground.2,4 No source opened for this entry explains the gap.
  • What water does to roasted bean texture: a 2011 study found that water first hardens green and roasted beans (antiplasticisation) and softens them only above a critical water activity9; a doctoral thesis calls the rise in firmness of quenched beans a plasticising effect of water.14 Both report firmer beans with more water at low moisture; they differ in the term.
Share or cite this entry

Change log

  • – The two conflicting glass transition studies are now described with their basis (2021 study, ground samples, midpoints, endpoints −1.7 to −2.5 °C; 2019 abstract silent on onset or midpoint and on sample form); the claim that both studies show water lowering Tg was replaced by the 2021 authors' own statement and the observation that their samples differed by less than 1 °C. Reason: Verification pass; the 2021 paper (Table 1) reports midpoints of −11.24, −11.93 and −11.87 °C with no significant difference, the green sample having the highest water activity.
  • – The 2007 texture study no longer says that water "above a critical level" removed brittleness; the region above the critical water activity is now described as progressive softening. Reason: Verification pass; the abstract ties loss of brittleness to plasticisation without placing it above or below the critical water activity, and reports softening above it.
  • – The roast-path reading of the state diagram and the link between rubbery state and swelling in Risks are now attributed to the textbook chapter; the earlier explanation of fast-roast expansion is identified as a patent reviewed in a thesis; the DMTA temperature (260 °C, thesis only) was removed. Reason: Verification pass; textbook-only and single Tier B claims (number rule).

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. 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
  3. ASablani, S. S.; Syamaladevi, R. M.; Swanson, B. G. (2010). A review of methods, data and applications of state diagrams of food systems. Food Engineering Reviews 2(3), 168–203 (author's submitted version, University of Alberta repository). doi:10.1007/s12393-010-9020-6 · other copies
  4. ABrzezińska, R.; Górska, A.; Wirkowska-Wojdyła, M.; Ostrowska-Ligęza, E.; Bryś, J. (2021). Thermal and Kinetic Properties of Brazilian Coffea Arabica Beans. Applied Sciences 11(14), 6324. doi:10.3390/app11146324 · other copies
  5. AFadai, N. T.; Please, C. P.; Van Gorder, R. A. (2019). Modelling structural deformations in a roasting coffee bean. International Journal of Non-Linear Mechanics 110, 123–130. doi:10.1016/j.ijnonlinmec.2018.12.006 · other copies
  6. ASchenker, S.; Handschin, S.; Frey, B.; Perren, R.; Escher, F. (2000). Pore structure of coffee beans affected by roasting conditions. Journal of Food Science 65(3), 452–457. doi:10.1111/j.1365-2621.2000.tb16026.x · other copies
  7. BSchenker, S. (2000). Investigations on the Hot Air Roasting of Coffee Beans (Diss. ETH No. 13620). ETH Zürich (doctoral thesis).
  8. BSmrke, S.; Sage, E.; Wellinger, M.; Yeretzian, C. (2018). The Coffee Freshness Handbook (ISBN 978-0-9995807-2-1). Specialty Coffee Association, Santa Ana.
  9. ARocculi, P.; Sacchetti, G.; Venturi, L.; Cremonini, M.; Dalla Rosa, M.; Pittia, P. (2011). Role of Water State and Mobility on the Antiplasticization of Green and Roasted Coffee Beans. Journal of Agricultural and Food Chemistry 59(15), 8265–8271. doi:10.1021/jf201333a · other copies
  10. APittia, P.; Nicoli, M. C.; Sacchetti, G. (2007). Effect of moisture and water activity on textural properties of raw and roasted coffee beans. Journal of Texture Studies 38(1), 116–134. doi:10.1111/j.1745-4603.2007.00089.x · other copies
  11. APittia, P.; Dalla Rosa, M.; Lerici, C. R. (2001). Textural changes of coffee beans as affected by roasting conditions. LWT – Food Science and Technology 34(3), 168–175. doi:10.1006/fstl.2000.0749 · other copies
  12. ABaggenstoss, J.; Poisson, L.; Luethi, R.; Perren, R.; Escher, F. (2007). Influence of Water Quench Cooling on Degassing and Aroma Stability of Roasted Coffee. Journal of Agricultural and Food Chemistry 55(16), 6685–6691. doi:10.1021/jf070338d · 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. BBaggenstoss, J. (2008). Coffee roasting and quenching technology – formation and stability of aroma compounds (Diss. ETH No. 17696). ETH Zürich (doctoral thesis). doi:10.3929/ethz-a-005666902 · other copies
  15. Bvon Blittersdorff, M.; Klatt, C. (2017). The Grind – Particles and Particularities. In: Folmer, B. (ed.), The Craft and Science of Coffee, Academic Press, ch. 13, pp. 311–328. doi:10.1016/B978-0-12-803520-7.00013-X · other copies

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 Glass–rubber transition. Tell us what you would change and why. A source that supports it helps most.