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

Roast profiles

The time–temperature curve of a roast – what the probes really measure, how rate of rise, turning point, first crack and development time are read, how roasters steer the curve, and what research says about profiles and flavour.

Across the chain: Science · Equipment

Source-checkedby Coffearium verification pass· Last source-checked 12 Oct 2026· Prepared with AI assistance· 17 sources (11 tier A)· 4 contested· Updated 12 Oct 2026
Reading depth
Contents
  • A roast profile is the course of temperature over time in a roast, logged from a probe in the bean mass; the reading is a mix of bean surface and hot air and is specific to the machine.1
  • The rate of rise is how fast that reading climbs per interval; in one study it was highest just after the turning point and lower at the end in every profile, though by very different amounts.4
  • The dip after charge and the turning point come from cold beans and air surrounding the probe, and from the probe’s own response time.5
  • In sensory studies, roast colour predicted flavour better than timing; at the same colour, development time after first crack mattered more than time to first crack.2

What a roast profile is#

A roast profile is the course of temperature over time during a roast. Roasters log the temperature from a probe in the bean mass at set time intervals, together with the time of first crack, from the charge (coffee entering the roaster) to the drop (coffee leaving it).1 The word is used in two senses: the planned path of the beans through the machine, and the flavour that path is meant to produce in the cup.1

A roaster steers four main parameters: the degree of roast (final product temperature), the roasting time, the shape of the time–temperature curve and the air-to-bean ratio.1 Two roasts can reach the same colour in the same total time by different paths, and the path changes how much water leaves the bean early, the reactions that follow and the final bean structure.1

Where the overview of roasting describes what happens inside the bean, this entry is about the curve itself: how it is measured, how it is read and how it is controlled.

What the probes measure#

Real bean surface or core temperature is hard to measure. Production roasters therefore use a probe placed where it stays in contact with the beans but also with hot air. Its reading is always a mix of bean surface and air temperature: good enough for process control, but specific to the machine design, so values from one roaster cannot simply be compared with another.1

  • Bean probe. The thermocouple in the bean mass. Its readings, logged over time, are the basis of the profile.1 Research roasters are sometimes fitted with extra sensors to follow bean temperature.11
  • Air probes. Roasters also log air temperatures. A pilot-scale air roaster, for example, recorded inlet air, outlet (exhaust) air and product temperature.5 Conventional automatic control holds a set hot-air temperature by adjusting the burner.1 A flame change takes longer to change the roasting environment than an airflow change.1
  • Bean core. Core temperature has been measured in small-scale experiments but is usually not possible in industrial roasting1; a fine thermocouple is pushed into a drilled hole. In a laboratory fluid bed, the bean core never reached the set air temperatures of 260 °C and 228 °C.6

The dip after charge. Before the charge, the roaster is preheated (in one study to 210 ± 5 °C for at least 30 min).4 When the green coffee goes in, the reading drops, because beans and air at ambient temperature (about 20 °C) surround the probe.5 The lowest point is the turning point; after it, the reading rises quickly because the temperature difference between the hot air and the cool beans is large.5 A thermocouple also takes time to respond, and one heat-transfer model adds a correction for exactly that.5

The dip and the turning point therefore describe the probe in its machine as much as the beans. The roasting overview’s warning applies with full force here: the reading depends on probe position, air and machine design, so a turning point is a useful marker for comparing batches on one roaster, not a bean temperature to compare between roasters.1,5

Rate of rise#

The rate of rise (RoR) is the slope of the bean-probe curve: how much the temperature climbs per unit of time. Mathematically it is the first derivative of temperature over time. One study logged it with roasting software as the temperature increase per 30 s.4

  • It peaks just after the turning point. In the same study, the “initial RoR” was defined as the highest positive RoR immediately after the turning point.4
  • It usually falls. As the beans heat up, the gap between air and bean narrows and so does the heat flow. In heat-transfer models, the heat flux at the bean surface is proportional to the difference between the roasting temperature and the bean surface temperature9, and in an air roaster at low load, the rate of change of product temperature tended to zero.5
  • But a falling RoR is a choice, not a law. In seven profiles of one coffee roasted on a 5 kg drum, the initial RoR was 5.26–10.37 °C per 30 s. The final RoR, over the last minute, ranged from 0.08 °C per 30 s in a “fast start” profile to 5.32 °C per 30 s in an “extended Maillard” profile.4
  • It reads the probe, not the bean. Because the RoR is computed from the bean-probe reading, it inherits that reading’s machine-specific mix of bean and air temperature1 and the thermocouple’s response time.5 How the interval and smoothing settings of logging software shape the RoR has not been studied in the sources found.

Roasters act on it: at first crack, for example, a roaster may adjust the flame and log the change to finesse the rate at which heat rises.1 Practitioner terms for particular RoR shapes – the “crash” and the “flick” – are explained in the Deep dive.

Landmarks on the curve#

Roasters read the curve through a handful of landmarks. The figure shows them on a schematic roast; the legend gives what the sources report for each.

A · TemperatureTimeBean probe12345678B · Rate of rise of the bean probeTimeRate of riseA · TemperatureTimeBean probe12345678B · Rate of rise of the bean probeTimeRate of rise

Bean-probe temperature (accent, A) and its rate of rise (B): one common shape that falls steadily (solid) and one that starts lower and stays higher to the drop (dashed)

  1. Charge. Green coffee enters the preheated roaster.1,4 Beans and air at ambient temperature surround the probe, and its reading falls.5
  2. Turning point. The lowest probe reading. After it the reading rises rapidly5, and the rate of rise is highest just afterwards.4 Comparable only on one machine and probe.1
  3. Drying. From charge until the beans visibly change colour (the "pre-colour change" phase in one study)4; generally the longest phase.9
  4. Colour change. The beans turn from greenish yellow to yellowish brown.4 A roaster logs it with time and probe temperature, together with the aroma from the trier.1
  5. First crack. Timed by ear, when a popping sound is first heard.2 Reported at 180–200 °C, without saying whether bean or air temperature14; the probe reading at that moment depends on the machine.1
  6. Development time. From first crack to the end of the roast.2,3 At a fixed colour, its length mattered more for flavour than the time to first crack.2
  7. Drop. The beans leave the roaster.1 At one colour (Agtron 76), roasts with longer development were ended at lower temperatures: 204.1 °C after 90 s, 191.1 °C after 390 s.3
  8. Cooling. The beans are cooled quickly with air or a water quench, which stops the roast.1
An annotated roast curve (A) with its rate of rise (B); the shapes show how the curves relate, not measured values. The two rate-of-rise traces are examples only: across seven profiles in one study, the final rate of rise ranged from 0.08 to 5.32 °C per 30 s.4 Schematic, not to scale.

Development time ratio. In the trade press, some roasters express development as a share of the total roast time, the development time ratio; this is a practitioner convention.17 The studies that tested development time worked in absolute seconds, not ratios.2,3 Converted from the timings of one study (first crack at about 570 s, development 90–390 s), its four roasts spanned roughly 14% to 41% (our calculation).3 No peer-reviewed test of the ratio itself was found.

Second crack. Lighter roasts are cooled soon after first crack, darker ones near second crack.9 Some research profiles run through it: in the titratable-acidity study above, all roasts lasted 16 min and continued past second crack.4

Logging and comparing profiles#

A roast log can be kept with a stopwatch, notebook and pencil, or on a laptop.1 It should hold at least the bean-probe temperature at set intervals and the time and temperature of first crack.1 Logging software records the curve and the RoR automatically; studies cited here used commercial roasting software for this.3,4 The senses still matter: if a probe fails or drifts out of calibration, a roaster who can read colour, size, texture and aroma from the trier can still finish the roast.1

Comparing profiles between machines is hard. The probe reading is specific to the machine design.1 Airflow, drum speed and the air-to-bean ratio change how heat reaches the beans.1 Studies define roast phases by events the beans show – colour change, first crack, second crack2,3,4 – and the sensory studies report final colour, not probe temperatures alone.2,3 When researchers did make the bean temperature follow the same curve on a drum and a laboratory fluid bed, the kinetics of aroma formation were similar in both.7

Reference values#

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

Parameter Value reported Source
What the production probe reads Mix of bean surface and hot air; specific to machine design Schenker & Rothgeb1
Laboratory bean-core thermocouple Type K, 0.25 mm, in a 0.3 mm drilled hole Geiger et al.6
Bean core vs set air temperature (laboratory fluid bed) Core stayed below 260 °C (HTST, 170 s) and 228 °C (LTLT, 720 s) Geiger et al.6
Cause of the dip after charge Beans and air at ambient temperature (about 20 °C) reach the probe Al-Shemmeri et al.5
Rate of rise, unit used °C per 30 s, logged with roasting software Anokye-Bempah et al.4
Initial RoR (highest after the turning point), 5 kg drum 5.07–10.37 °C per 30 s across 13 roasts of three coffees; slow-start 5.07–5.33, fast-start 9.91–10.15 Anokye-Bempah et al.4
Final RoR (last minute), same roasts 0.01–5.43 °C per 30 s Anokye-Bempah et al.4
Roaster preheating, same study 210 ± 5 °C for at least 30 min; 4 kg charged in a 5 kg drum Anokye-Bempah et al.4
First crack, same study (16 min roasts) About 8 min (fast start); 12 min (slow start) Anokye-Bempah et al.4
First crack, development-time study (1 kg drum) About 570 s Alstrup et al.3
Development time at one colour (Agtron 76 ± 1) 90 s, 143 s, 266 s, 390 s Alstrup et al.3
End-of-roast temperature for those roasts 204.1 °C, 201.2 °C, 198.4 °C, 191.1 °C Alstrup et al.3
Development as share of total time, those roasts About 14%, 20%, 32%, 41% (our calculation) from Alstrup et al.3
Roast times, eight sensory studies 7:40–20:20 min total; development 1:20–6:45 min (0:10 in one excluded study) Münchow et al.2
Correlation with flavour attributes (mean absolute r) Study 1: colour 0.40, development 0.21, time to first crack 0.09 (n.s.); Study 2: colour 0.46, development 0.07 (n.s.) Münchow et al.2
Time after first crack, drum-roaster study 1.0 and 2.0 min compared; 30 s was enough to lower cup quality by a couple of points Catão et al.12
Underdeveloped defect roast Air temperature at charge 135 °C (normal roast 210 °C); 8 min longer than normal Giacalone et al.11
First crack, temperature (bean or air not stated) 180–200 °C Bicho et al.14

Why the curve bends#

Heat flows down a temperature difference. Single-bean models treat the heat flux into the bean as proportional to the difference between the roasting temperature and the bean surface.9 At charge the difference is largest, which is why the reading climbs fast after the turning point5; as the bean warms, the difference and the heat flow shrink9, and how the rate of rise develops then depends on how the roaster sets the heat.4

Water slows the early roast. Evaporating water absorbs heat: in a 2025 single-bean model, evaporative cooling starts at 100 °C, and without heat generation the model underestimated bean temperature.10 Early in the roast a lot of energy goes into water; later the process becomes progressively more exothermic and needs much less heat input.2 The same 2025 model assumed, from earlier literature, heat generation inside the bean above 150 °C10; where exactly the switch happens is disputed (see the roasting overview).

Inside the bean. A 2017 study compared two single-bean models. The one adapted from earlier work predicts that after the first minute, temperature and moisture inside the bean become nearly uniform; the authors’ new model, whose drying front they consider more realistic, shows a much slower drying front.9 In extreme fast roasting, however, a steep gradient can leave the surface over-roasted and the core under-roasted.1 No probe in the bean mass can see this.

Probe lag. A thermocouple in the roasting chamber responds slowly. A spouted-bed model added a thermal response coefficient to correct for it, and the authors note that using the thermocouple reading as bean temperature reduces the accuracy of the fitted parameters.5

Steering the curve#

Air-temperature control versus product-temperature control. Traditional automation sets and holds a hot-air temperature, in one stage or several. The beans’ temperature then varies from batch to batch with green-coffee moisture, weather and a cold roaster. More advanced systems follow a product-temperature master curve stored in the recipe and adjust the energy input continuously, so the beans see the same temperature development every batch.1

Manual levers. On small drum roasters the artisan works mainly with the flame and the trier; drum speed and airflow are used less often. Airflow changes convection almost instantly, whereas a flame change takes longer to reach the beans. Too high a drum speed can press beans against the hot wall and scorch the flats.1 A specialty roast typically starts near maximum burner output and then steps the power down to slow the browning reactions.3

In a multivariate study on a professional roaster, the variables tested were initial drum temperature, the airflow opening, drum speed, the rate of increase in gas pressure to the burners, and the time after first crack; time after first crack had the largest influence on both sensory and chemical data.12

Shapes of the curve. Keeping final colour and total time constant, a slow-start-fast-final profile usually gave more bean volume and porosity than a traditional one, sometimes a higher extraction yield, and a statistically different flavour profile.1 The 2024 titratable-acidity study compared seven profiles named fast start, slow start, medium, production, exaggerated flick, negative rate of rise and extended Maillard; fast start meant high initial heat followed by a constant deceleration, slow start the reverse.4

024681012°C per 30 sFilled: initial (just after turning point) · open: final (last minute)Fast start10.154 0.084 Medium9.114 0.634 Negative rate of rise10.114 0.324 Production8.194 3.694 Exaggerated flick10.374 4.934 Extended Maillard10.204 5.324 Slow start5.26 · 4.564
Data table
Point°C per 30 sSource
Fast start10.1510.154
Fast start0.080.084
Medium9.119.114
Medium0.630.634
Negative rate of rise10.1110.114
Negative rate of rise0.320.324
Production8.198.194
Production3.693.694
Exaggerated flick10.3710.374
Exaggerated flick4.934.934
Extended Maillard10.2010.24
Extended Maillard5.325.324
Slow start5.265.264
Slow start4.564.564
Initial rate of rise (highest just after the turning point) and final rate of rise (last minute) for seven profiles of one washed African coffee roasted for 16 min in a 5 kg drum.4

Crash and flick#

Two RoR shapes have practitioner names. In practitioner usage, a crash is a sudden plunge of the bean-temperature RoR during first crack; a flick is the RoR recovering and turning upwards again towards the end of first crack or the end of the roast.15,16 One practitioner course claims that a crash can make a coffee slightly less round and sweet and that a flick leads to roasty or ashy flavours.15

The evidence is thin. No peer-reviewed sensory study of crashes or flicks was found. The one peer-reviewed study that roasted an “exaggerated flick” profile measured titratable acidity, not flavour.4 Neither the mechanism of the crash nor whether it reflects the beans or the probe has been established in the sources found; the probe reading is shaped by heat released inside the beans, evaporative cooling and the thermocouple’s response time.5,10

Profiles and flavour: what research shows#

Colour first. Eight sensory studies on a 1 kg drum and a 15 kg Loring roaster found that, in the two studies that varied both, roast colour predicted flavour better than timing.2 The authors note that colour measurement is often left out of quality control in specialty roasteries.2

Then development time. At constant colour, development time had a larger effect on flavour than time to first crack.2 In a follow-up at one colour, development of 90–390 s shifted the cup from fruity and acidic towards roasty and bitter, and the acids measured generally declined with longer development.3 To keep the colour the same, the end temperature had to be lowered as development got longer.3

185190195200205210°C at the end of the roastFast · 90 s development204.1 °C3 Medium · 143 s201.2 °C3 Slow · 266 s198.4 °C3 Baked · 390 s191.1 °C3
Data table
Point°C at the end of the roastSource
Fast · 90 s development204.1 °C204.13
Medium · 143 s201.2 °C201.23
Slow · 266 s198.4 °C198.43
Baked · 390 s191.1 °C191.13
End-of-roast temperature of four roasts taken to the same colour (Agtron 76) with different development times after a first crack at about 570 s, 1 kg drum roaster.3

Path matters at fixed colour. On the same raw material and roaster, roasting to the same degree along different time–temperature paths gives different flavour profiles.8 High-temperature short-time roasting changed bean physical properties and the kinetics of aroma formation compared with low-temperature long-time roasting7; a doctoral thesis on hot-air roasting called the temperature profile the most crucial parameter in process design.13

Acidity over the roast. In a 5 kg drum, titratable acidity rose from the start of the roast, peaked at first crack whatever the profile, and had fallen back almost to its initial value by the onset of second crack. The profile changed how fast this happened, but the peak itself differed little between profiles.4

Same curve, different machine. When a drum and a laboratory fluid bed followed the same bean-temperature profile, aroma formation was similar.7,8 The catch for practice is that production probes do not show the bean temperature directly.1

Risks and defects#

  • Stalling and baking. A “baked” roast stretches the time after first crack with little or no rise in temperature.3 On the curve this is a flat line with a rate of rise near zero.
  • Underdevelopment. In a defect study, an underdeveloped roast was made with a much lower air temperature at charge (135 °C) and 8 min more roast time; the stalling early in the curve was meant to suppress typical coffee aromas, yet most aroma compounds ended up at levels comparable to the normal roast.11
  • Running too fast. Very fast heating can over-roast the surface while the core stays under-roasted, giving burnt and greenish notes together.1
  • Trusting the probe too far. Probes can fail or drift out of calibration; opening the trier also lets cool air into the drum and can change the final stages of the roast.1
  • Copying profiles between roasters. Because readings are machine-specific, a profile copied as numbers from another roaster cannot be expected to reproduce the same bean temperatures.1

Open questions#

  • RoR shapes and flavour. No peer-reviewed sensory test of crashes, flicks or a declining RoR as such was found; the one study using these profiles measured acidity only.4
  • Development time ratio. The ratio is promoted in the trade press17, but the studies found tested development time in seconds.2,3
  • From probe to bean. Models can estimate bean temperature from operating conditions5,10, but production roasters still log a machine-specific probe reading.1
  • Software settings. How the interval and smoothing chosen in logging software change the displayed RoR, and the decisions roasters take from it, has not been studied in the sources found.
  • Range of coffees. The sensory evidence comes mainly from washed Arabica on a few drum and convection roasters.2

Key studies#

Münchow et al., 2020 – Colour versus timing. Pools eight descriptive sensory studies of washed Arabica on a 1 kg drum and a 15 kg Loring roaster. Colour was the stronger predictor of flavour; at constant colour, development time after first crack mattered more than time to first crack.

Beverages 6(2), 29 · doi:10.3390/beverages6020029

Alstrup et al., 2020 – Development time at one colour. Four roasts of one Colombian coffee to Agtron 76, similar up to first crack at about 570 s and then developed for 90–390 s, logged with roasting software. Sensory panels, GC–MS and NMR all separated the roasts, and longer development needed a lower end temperature for the same colour.

Beverages 6(4), 70 · doi:10.3390/beverages6040070

Anokye-Bempah et al., 2024 – Profiles and titratable acidity. Up to seven roast profiles on a 5 kg drum, sampled every minute of 16 min roasts of three coffees. Titratable acidity always peaked at first crack; the profile changed the timing of the rise and fall more than the peak.

Scientific Reports 14, 8237 · doi:10.1038/s41598-024-57256-y

Baggenstoss et al., 2008 – Time–temperature conditions and aroma. Follows 16 aroma compounds in a commercial drum and laboratory fluid beds, roasting fast and slow to the same lightness. Fast roasting changed bean properties and aroma kinetics; when both roasters followed the same bean-temperature profile, the kinetics of aroma formation were similar.

Journal of Agricultural and Food Chemistry 56(14), 5836–5846 · doi:10.1021/jf800327j

Al-Shemmeri et al., 2025 – What the thermocouple sees. Models heat and mass transfer in a 0.5 kg spouted-bed roaster from logged inlet, outlet and product temperatures. Explains the dip after charge and the turning point, and corrects for the thermocouple’s slow response.

Beverages 11(6), 162 · doi:10.3390/beverages11060162

Contested
  • Does the same bean-temperature curve give the same coffee on different machines? When one bean-temperature profile was applied in a laboratory fluid bed and a drum roaster, the kinetics of aroma formation were similar7,8; but production probes read a machine-specific mix of bean and air temperature1, and changing only the air-to-bean ratio made a limited but statistically significant difference in flavour.1
  • Fast or slow roasting and acidity: a textbook chapter says fast-roasted coffee usually shows more acidity and slow-roasted coffee often more "fruity" notes1; in a multi-study sensory analysis, longer roasting and development went with less fruitiness and acidity2; in a 5 kg drum, the profile strongly changed how titratable acidity rose and fell, but the peak hardly differed between profiles.4
  • What "development" means: two sensory studies define development time as first crack to the end of the roast2,3; a titratable-acidity study calls the span from first crack to second crack the development phase.4
  • Crash and flick: practitioner sources say a flick in the rate of rise late in the roast leads to roasty, even ashy flavours15; no peer-reviewed sensory test of this was found. One study roasted an "exaggerated flick" profile but measured only titratable acidity.4
Share or cite this entry

Change log

  • – The sentence on temperature and moisture inside the bean now says that the near-uniform result comes from the simpler of two models in the 2017 study, and that the authors consider their second model, with a slower drying front, more realistic. Reason: Cross-check during verification of the new drying-phase entry; Fadai et al. 2017 (source 9) state that Model 2 has the more realistic drying front.

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. 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
  5. AAl-Shemmeri, M.; Fryer, P. J.; Farr, R.; Lopez-Quiroga, E. (2025). Batch-Scale Simulation of Heat and Mass Transfer of Coffee Roasting in Spouted Bed Roasters. Beverages 11(6), 162. doi:10.3390/beverages11060162 · other copies
  6. 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
  7. ABaggenstoss, J.; Poisson, L.; Kaegi, R.; Perren, R.; Escher, F. (2008). Coffee Roasting and Aroma Formation: Application of Different Time−Temperature Conditions. Journal of Agricultural and Food Chemistry 56(14), 5836–5846. doi:10.1021/jf800327j · other copies
  8. 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
  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. 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
  11. AGiacalone, D.; Degn, T. K.; Yang, N.; Liu, C.; Fisk, I.; Münchow, M. (2019). Common roasting defects in coffee: Aroma composition, sensory characterization and consumer perception. Food Quality and Preference 71, 463–474. doi:10.1016/j.foodqual.2018.03.009 · other copies
  12. ACatão, A. A.; Mateus, N. de S.; Garcia, C. da C.; da Silva, M. C.; Novaes, F. J. M.; Alves, S.; Marriott, P. J.; da Silva, A. I. (2022). Coffee-roasting variables associated with volatile organic profiles and sensory evaluation using multivariate analysis. Applied Food Research 2(2), 100223. doi:10.1016/j.afres.2022.100223 · other copies
  13. BSchenker, S. (2000). Investigations on the Hot Air Roasting of Coffee Beans (Diss. ETH No. 13620). ETH Zürich (doctoral thesis).
  14. ABicho, N. C.; Leitão, A. E.; Ramalho, J. C.; Lidon, F. C. (2012). Use of colour parameters for roasted coffee assessment. Ciência e Tecnologia de Alimentos 32(3), 436–442. doi:10.1590/S0101-20612012005000068 · other copies
  15. CBarista Hustle HTR 3.04 Fixing Crashes (How to Roast course). Barista Hustle (online coffee education).
  16. CCropster (2021). Cropster Adds Flick Prediction To AI Features (press release). Sprudge.
  17. CRao, S. (2014). Roast Development Time and the Prerequisite for Any Successful Roast. Daily Coffee News.

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 Roast profiles. Tell us what you would change and why. A source that supports it helps most.