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

Endothermic and exothermic phases

Why a roast first soaks up heat and later starts to release heat of its own – what calorimetry and models show, why nobody agrees where the switch happens, and what it means for heat control and cooling.

Across the chain: Science

Source-checkedby Coffearium verification pass· Last source-checked 19 Oct 2026· Prepared with AI assistance· 15 sources (12 tier A)· 2 contested· Updated 19 Oct 2026
Reading depth
Contents
  • Early in the roast, heat goes mainly into warming the beans and evaporating water (endothermic); later the process becomes progressively more exothermic and needs much less heat input.1,2
  • Where the switch happens is disputed; sources place it at first crack, before it, from 150 °C, near 200 °C or around 250 °C.1,13,7,6,9
  • In a 2025 bean model, water evaporation affected bean temperature more than reaction heat; both terms were needed to fit the data.7
  • Thermal analysis of dried, ground green coffee in argon showed mainly heat uptake between 100 and 245 °C and heat release only above 245 °C.10
  • According to a roasting textbook chapter, the late, heat-releasing stage speeds up, so the roast must be stopped abruptly and cooled efficiently.1

What the two words mean#

Endothermic means that heat is taken up; exothermic means that heat is released.1 In a roast, both kinds of process run side by side, and the balance between them shifts as the roast goes on.1,7

At the start, a large energy input is needed to evaporate water and to set chemical reactions going. A roasting textbook chapter calls this the endothermic phase.1 Towards the end, much less heat input is needed, because the process itself becomes progressively more exothermic.2

A second chapter of the same textbook describes the roast in the same way: an endothermic drying phase, then an exothermic phase in which most flavour compounds form, then cooling.15 The glossary terms endothermic reaction and exothermic reaction give the short definitions.

Heat in, heat out#

During a roast, heat flows into the bean from hot air and hot surfaces. The heat transfer entry describes those paths. Inside the bean, the heat has three main uses or sources:

  • Warming. Heat raises the temperature of the solid bean and the water in it.2
  • Evaporation. Turning water into vapour takes up a large amount of heat. One model uses 2.3 million joules per kilogram of water.6 In a 2025 model, this evaporative cooling starts at 100 °C.7
  • Reactions. Some reactions take up heat and others release it. According to the textbook chapter, the beans eventually start to generate heat on their own.1

Early on, heating the water dominates, so the roast is primarily endothermic.2 Later, the process becomes progressively more exothermic, and much less heat input is needed.2 In a 2025 single-bean model, evaporation still had the larger effect on bean temperature than reaction heat.7

1Early roast: heat taken upHeat warms the beans and their water 2 Evaporation absorbs heat 7 2Transition (disputed)Before first crack (off-gas) 13 At first crack, in a textbook 1 150 to about 250 °C (models) 7,9 3Late roast: heat releasedReactions release heat of their own 1,7 Much less heat input needed 2 4Drop and coolingStop abruptly and cool fast 1,4 Otherwise the roast runs on 1 heat in: high burner output 3 heat in: stepped down 3 heat out: cooling 1 course of the roastheat flow (direction only, not amount)Schematic, not to scale. Where the transition lies is disputed.
The heat balance of a roast in four steps: heat uptake early, a disputed transition, heat release late, and an abrupt stop.1,2,3,4,7,9,13 Schematic, not to scale; the arrows show direction, not amounts.

Where the balance turns#

The sources do not agree on when the roast turns exothermic. A roasting textbook chapter ties the switch to first crack.1 A study of aroma compounds in the exhaust of a production drum roaster placed it at 625 s, before first crack at about 730 s.13

Model studies give temperatures instead. One assumes reaction heat from 150 °C.7 Another says exothermic reactions can begin once the bean nears 200 °C.6 A 2013 study links them to roasting temperatures above 230 °C, and a 2009 study places them at about 250 °C.8,9

Laboratory heat measurements disagree too. In one, ground green coffee mainly took up heat up to 245 °C.10 In another, six green coffees each switched at a different temperature.11 Part of the spread comes from method: the studies measure a whole bean, a model, a ground sample or the exhaust gas.

100120140160180200220240260280°CReaction CO₂ and waterbegin(bean core, fluid bed)180–200 °C5 Reaction heat switched on(2025 bean model)above 150 °C7 Exothermic reactions canbegin(2017 model)nearly 200 °C6 Exothermic reactions(2013 CFD; air, not bean)above 230 °C8 Beans burn, model deviates(2009 model)about 250 °C9 Net heat uptake(dried, ground green coffee)100–245 °C10 Heat release begins(same study)from 245 °C10
Data table
Point°CSource
Reaction CO₂ and water begin (bean core, fluid bed)180–200 °C180–2005
Reaction heat switched on (2025 bean model)above 150 °C1507
Exothermic reactions can begin (2017 model)nearly 200 °C2006
Exothermic reactions (2013 CFD; air, not bean)above 230 °C2308
Beans burn, model deviates (2009 model)about 250 °C2509
Net heat uptake (dried, ground green coffee)100–245 °C100–24510
Heat release begins (same study)from 245 °C24510
Temperatures reported for the start of reaction gases and of heat-releasing reactions, by method. Values are not comparable one to one: they come from a bean core, models, a dried ground sample and different roasting temperatures.5,6,7,8,9,10

What it means at the roaster#

Heat input goes down. A 2020 study describes the usual practice: start near maximum burner output, then step it down as the roast goes on.3 The stated aim is to slow the browning reactions.3 A roasting textbook chapter adds that a roaster may adjust the flame at first crack, linking this to the switch to exothermic.1

The roast gains momentum. According to the textbook chapter, the final stages speed up and gradually approach the conditions of a combustion process.1 A few seconds can then separate the intended roast degree from an over-roasted coffee.1

The stop must be abrupt. For this reason, the chapter calls for an abrupt stop with efficient precooling or cooling.1 A doctoral thesis also states that the roast must be stopped by rapid cooling.4 The cooling and quenching entry describes how.

Fire. According to the same chapter, beans can catch fire if the roast runs on uncontrolled.1 A 2009 model study notes that exothermic reactions near 250 °C lead the beans to burn, but that industrial roasts normally stop well before that point.9

Reference values#

Values reported in the sources. They come from different methods, samples and bases, and are not targets.

Parameter Value reported Method and basis Source
Reaction CO₂ and water begin 180–200 °C Thermocouple in the bean core; laboratory fluid bed, 100 g batches Geiger et al. 20055
Exothermic reactions can begin Nearly 200 °C Statement in a heat-and-mass-transfer model study Fadai et al. 20176
Reaction heat switched on Above 150 °C Lumped single-bean model; threshold taken from a 2002 model Bustos-Vanegas et al. 20257
Total reaction heat in models 232 kJ per kg dry matter Literature parameter used in two model studies (per kg dry matter in the 2025 paper; the 2013 paper gives kJ/kg) Bustos-Vanegas et al. 20257; Alonso-Torres et al. 20138
Exothermic reactions occur At roasting temperatures above 230 °C (roasting, not bean, temperature) Single-bean CFD study; reaction heat not in its model Alonso-Torres et al. 20138
Exothermic reactions, beans burn About 250 °C Lumped model compared with published bean data; better at 190 °C than at 290 °C air (values in the text; the figure captions give 180 and 300 °C) Basile & Kikic 20099
Net heat uptake 100–245 °C; about 100 kJ/kg Thermal analysis, 2 g dried, ground green coffee, argon, 20 °C/min Strezov & Evans 200510
Heat release From 245 °C; about −60 and −20 kJ/kg in two reactions Same study, up to 500 °C Strezov & Evans 200510
Endothermic then exothermic, switch differs by coffee Six green coffees, 20–250 °C DSC, 10 mg cross-sections, nitrogen, 10 °C/min Mutovkina & Bredikhin 202311
Transition endothermic to exothermic 625 s (first crack about 730 s) Off-gas aroma compounds (real-time PCA), one dark roast of 15.6 kg, gas-heated drum Wieland et al. 201213
Latent heat of water in models 2.3 × 10⁶ J/kg Model parameter Fadai et al. 20176
Water evaporated during roasting 10.2 g (fast) and 11.4 g (slow) per 100 g green coffee Exhaust gas analysis, laboratory fluid bed Geiger et al. 20055

Thermodynamics of the roast#

Three terms. A single-bean energy balance has a heat-transfer term from the surroundings, an endothermic evaporation term proportional to the loss of moisture, and an exothermic reaction term.7 In a 2025 model, the reaction term was switched on above 150 °C. Its rate followed an Arrhenius law and stopped once the total heat of reaction was reached.7 That total was 232 kJ per kg dry matter, a value both this study and a 2013 study take from a 2002 model.7,8 The two papers print different values for the activation parameter of that model (5500 K and 500 K).7,8

Evaporation against reaction heat. In the 2025 model, leaving out evaporation made the model overestimate bean temperature during the exothermic period from 150 °C. Leaving out reaction heat made it underestimate.7 Evaporation had the stronger effect, but both terms were needed for the best fit.7

Orders of magnitude. In a fluid-bed study, 10.2–11.4 g water per 100 g green coffee evaporated during roasting.5 At 2.3 × 10⁶ J/kg6, that is about 235–262 kJ per kg green coffee (our calculation). This is the same order as the 232 kJ per kg dry matter of reaction heat used in the models.7 Part of the evaporated water is itself formed by reactions, so the comparison is rough.5

What the reactions are. According to the textbook chapter, Maillard reaction, Strecker degradation, pyrolysis and caramelisation are among the main reactions of carbohydrates during roasting.1 The off-gas study links the exothermic phase to Maillard reactions forming coffee aroma compounds.13 The 2025 model describes the heat as coming from oxidation reactions inside the grain.7 No source opened for this entry measured the heat of single roasting reactions inside a bean.

050100150200250300kJ per kg (basis in the label)Taken up, 100–245 °C(thermal analysis, dried groundgreen coffee)100 kJ/kg10 Released, first exotherm(same study)60 kJ/kg10 Released, second exotherm(same study)20 kJ/kg10 Reaction heat used inmodels(per kg dry matter)232 kJ/kg7 Evaporating the waterlost, fast roast(per kg green coffee; ourcalculation)≈235 kJ/kg5,6 Evaporating the waterlost, slow roast(per kg green coffee; ourcalculation)≈262 kJ/kg5,6
Data table
kJ per kg (basis in the label)Source
Taken up, 100–245 °C (thermal analysis, dried ground green coffee)100 kJ/kg10
Released, first exotherm (same study)60 kJ/kg10
Released, second exotherm (same study)20 kJ/kg10
Reaction heat used in models (per kg dry matter)232 kJ/kg7
Evaporating the water lost, fast roast (per kg green coffee; our calculation)≈235 kJ/kg5, 6
Evaporating the water lost, slow roast (per kg green coffee; our calculation)≈262 kJ/kg5, 6
Heat quantities reported or derived, on their own bases: heat uptake and release in thermal analysis of dried, ground green coffee; reaction heat used in models; and heat to evaporate the water lost in two fluid-bed roasts (our calculation).5,6,7,10

Measuring heat release#

Differential scanning calorimetry (DSC). A DSC measures the heat flow into a small sample against an empty reference while both are heated at a set rate. A 2023 study heated 10 mg cross-sections of six green coffees from 20 to 250 °C at 10 °C/min under nitrogen.11 All showed an endothermic phase followed by an exothermic one, at a different temperature for each sample.11 The authors conclude that denser, moister coffees needed more heat.11 The paper’s table and text label the same values differently, so we do not give its peak temperatures here.11

Thermal analysis of ground coffee. A 2005 study heated 2 g of ground green coffee, dried under vacuum, in a silica tube under argon at 20 °C/min.10 It found mainly endothermic reactions between 100 and 245 °C, then exothermic decomposition with two reactions up to 430 °C.10 The heat taken up was estimated at 100 kJ/kg; the two exotherms released about 60 and 20 kJ/kg.10 Because the sample was dried first, the thermograms showed no sign of the bean’s own water.10

Atmosphere matters. In a 2005 thermoanalytical study in synthetic air, green coffee showed a broad endotherm from dehydration, followed by two exotherms.12 In freeze-dried samples, a strong exotherm above 360 °C was attributed to heat generated by the sample in an oxidising atmosphere.12 A roaster’s air is neither pure air nor inert gas, so these results transfer only in part.

Calorimetry of self-heating. The textbook chapter bases its statement on heat release on a 1982 calorimetric study of self-heating in coffee and chicory.1 That study could not be opened for this entry. A doctoral thesis lists earlier calorimetric and textbook values for the switch to exothermic, which disagree with one another.4

Fitting bean temperatures. Model studies compare a heat balance with measured bean temperatures; reaction heat is either fitted as a term or invoked to explain where the model deviates.7,9 In a 2009 study, a lumped model matched published bean data better at 190 °C air than at 290 °C. The authors attribute the deviation to exothermic reactions at about 250 °C.9

Off-gas signals. In a production drum, aroma compounds in the exhaust changed abruptly at 625 s and again at 730 s, with first crack.13 The authors read the first change as the transition to the exothermic phase. It was a change in the gas profile, not a heat measurement.13

Energy balance on a roaster. No study opened for this entry measured reaction heat from an energy balance over a production roaster.

How models handle reaction heat#

Model studies treat reaction heat in three ways:

  • Included as a source term. The 2025 single-bean model adds reaction heat above 150 °C and fits the data best with both evaporation and reaction heat.7
  • Left out. A 2013 CFD study of single beans lists the reaction-heat parameters but states that the major chemical transformations were not yet in its model.8 A 2017 heat-and-mass-transfer study treated its reaction-energy term as small and did not use it in a simplified model.6
  • Lumped into heat capacity. A 2009 model folds evaporation and other thermal effects into one apparent specific heat that depends on temperature and initial moisture.9

Relation to first and second crack#

First crack. In a fluid-bed study, reaction gases rose significantly only from a bean core temperature of 180–200 °C.5 The first crack entry covers the cracking itself and the temperatures reported for it. A textbook chapter ties the switch to exothermic to first crack.1 The off-gas study placed the transition about 105 s before first crack (our calculation from 625 s and 730 s).13 No source opened measured heat release in a bean at the moment of first crack.

Rate of rise. A 2024 study describes the practitioner term “flick”, a sudden increase in the rate of rise after first crack, as typically associated with the onset of exothermic reactions.14 In that study, the flick profile was made by changing the roaster’s gas and airflow, so it does not separate burner heat from reaction heat.14 The thesis found that published temperature curves do not show a late rise clearly attributable to exothermic reactions, and its own laboratory and industrial roasts showed no substantial additional rise either.4

Second crack. No source opened for this entry links second crack to a measured change in heat release. The strongest heat release in thermal analysis began at 245 °C.10 In one drum study, medium roasts ended at 191.1–204.1 °C (measuring point not stated).3

Risks#

  • Overshoot. In the late, exothermic stage, a few seconds can turn the intended roast degree into an over-roasted product.1 See roast defects.
  • Burning and fire. Beans can catch fire in an uncontrolled roast, according to the textbook chapter1; a model study places burning at about 250 °C.9
  • Late stop. The roast must be stopped abruptly and cooled efficiently, or it runs on after the drop.1,4
  • Copied set points. Temperatures for the switch come from different methods and bases; none is a probe value for a given roaster.5,7,10

Open questions#

  • Onset of net heat release. Reported positions range from 625 s in one drum roast to about 250 °C in a model (see Contested).13,9
  • Size of reaction heat in a whole bean. The models use one literature value7; the thermal analysis opened measured a dried, ground sample.10
  • Self-heating in roast curves. Whether the late roast shows a temperature rise from reactions alone is not settled.1,4
  • Coffee to coffee. One DSC study found the switch at different temperatures for six coffees11; how origin, processing and density shift it is not established.

Key studies#

Bustos-Vanegas et al., 2025 – Effect of heat generation. Single arabica beans were held in a gas-fired drum at 200–280 °C and their temperature fitted with a lumped model. The model included evaporation and a reaction-heat term from 150 °C; evaporation mattered more, but both were needed for the best fit.

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

Strezov & Evans, 2005 – Thermal analysis of green coffee. Dried, ground green coffee was heated under argon to 500 °C while heat flow was measured. Reactions were mainly endothermic from 100 to 245 °C, and two exothermic reactions followed at higher temperatures.

International Journal of Food Properties 8(1), 101–111 · doi:10.1081/JFP-200048060

Mutovkina & Bredikhin, 2023 – DSC of six green coffees. Cross-sections of six green coffees from four countries were scanned from 20 to 250 °C under nitrogen. All showed an endothermic phase followed by an exothermic one, at a different temperature for each coffee.

Food Science and Technology (Campinas) 43, e119722 · doi:10.1590/fst.119722

Wieland et al., 2012 – Off-gas monitoring of a drum roast. Aroma compounds in the exhaust of a 15.6 kg gas-heated drum roast were followed in real time. Two abrupt changes appeared, at 625 s and with first crack at 730 s; the authors read the first as the change from endothermic drying to the exothermic phase.

Analytical and Bioanalytical Chemistry 402(8), 2531–2543 · doi:10.1007/s00216-011-5401-9

Contested
  • Where the roast turns exothermic: a roasting textbook chapter ties the switch to first crack1; a study of off-gas aroma compounds in a 15.6 kg drum roast placed it at 625 s, before first crack at about 730 s13; a 2025 model assumed reaction heat above 150 °C7; a heat-transfer model says exothermic reactions can begin near 200 °C6; a 2013 model study links exothermic reactions to roasting temperatures above 230 °C8; a 2009 model study puts them at about 250 °C9; thermal analysis of dried, ground green coffee found heat release only above 245 °C10; and calorimetry of six green coffees found a different switch temperature for each.11 The sources measure different things (bean core, modelled bean, ground sample, off-gas) under different conditions.
  • Whether self-heating shows in roast curves: a roasting textbook chapter, citing a 1982 calorimetric study, states that the beans eventually start to generate heat on their own1; a doctoral thesis by the chapter’s first author found that published bean temperature curves do not show a late rise that can be clearly attributed to exothermic reactions, and its own laboratory and industrial roasts showed no substantial additional rise either.4
Share or cite this entry

Change log

  • – Removed the statement that 180–200 °C (start of reaction gases) is the range in which bean cracking is reported; the summary point on onset temperatures now cites the 2009 model study for about 250 °C instead of the 2013 study. Reason: Verification pass; Geiger et al. 2005 does not mention cracking, and the 250 °C value comes from Basile & Kikic 2009, not Alonso-Torres et al. 2013 (230 °C).
  • – Strezov & Evans 2005 method corrected (silica tube; the dried sample showed no sign of occluded water, as the authors state); Basile & Kikic 2009 described as comparing, not fitting, a model with published bean data, with the 190/290 °C versus 180/300 °C inconsistency of the source noted; the thesis's own finding of no substantial late temperature rise added. Reason: Verification pass against the full texts.

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. 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
  3. AAlstrup, J.; Petersen, M. A.; Larsen, F. H.; Münchow, M. (2020). The Effect of Roast Development Time Modulations on the Sensory Profile and Chemical Composition of the Coffee Brew as Measured by NMR and DHS-GC–MS. Beverages 6(4), 70. doi:10.3390/beverages6040070 · other copies
  4. BSchenker, S. (2000). Investigations on the Hot Air Roasting of Coffee Beans (Diss. ETH No. 13620). ETH Zürich (doctoral thesis).
  5. 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
  6. 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
  7. 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
  8. AAlonso-Torres, B.; Hernández-Pérez, J. A.; Sierra-Espinoza, F.; Schenker, S.; Yeretzian, C. (2013). Modeling and Validation of Heat and Mass Transfer in Individual Coffee Beans during the Coffee Roasting Process Using Computational Fluid Dynamics (CFD). CHIMIA 67(4), 291–294. doi:10.2533/chimia.2013.291 · other copies
  9. ABasile, M.; Kikic, I. (2009). A Lumped Specific Heat Capacity Approach for Predicting the Non-stationary Thermal Profile of Coffee During Roasting. Chemical and Biochemical Engineering Quarterly 23(2), 167–177.
  10. AStrezov, V.; Evans, T. J. (2005). Thermal Analysis of the Reactions and Kinetics of Green Coffee During Roasting. International Journal of Food Properties 8, 101–111. doi:10.1081/JFP-200048060 · other copies
  11. AMutovkina, E. A.; Bredikhin, S. A. (2023). Analysis of coffee thermophysical changes during roasting using differential scanning colorimetry. Food Science and Technology (Campinas) 43, e119722. doi:10.1590/fst.119722 · other copies
  12. APereira, C. A.; Schnitzler, E.; Carvalho Filho, M. A. S. (2005). Estudo termoanalítico (TG, DTG e DSC) dos cafés in natura e processados. Publicatio UEPG: Ciências Exatas e da Terra, Ciências Agrárias e Engenharias 11(1), 61–66.
  13. AWieland, F.; Gloess, A. N.; Keller, M.; Wetzel, A.; Schenker, S.; Yeretzian, C. (2012). Online monitoring of coffee roasting by proton transfer reaction time-of-flight mass spectrometry (PTR-ToF-MS): towards a real-time process control for a consistent roast profile. Analytical and Bioanalytical Chemistry 402(8), 2531–2543. doi:10.1007/s00216-011-5401-9 · other copies
  14. AAnokye-Bempah, L.; Styczynski, T.; de Andrade Teixeira Fernandes, N.; Gervay-Hague, J.; Ristenpart, W. D.; Donis-González, I. R. (2024). The effect of roast profiles on the dynamics of titratable acidity during coffee roasting. Scientific Reports 14, 8237. doi:10.1038/s41598-024-57256-y · other copies
  15. 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

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 Endothermic and exothermic phases. Tell us what you would change and why. A source that supports it helps most.