Roast defects
Flaws the roast itself puts into coffee – underdeveloped, baked, over-roasted, scorched, tipped, faced, cratered or uneven beans – how each looks, what causes it in the roaster, how it affects flavour and how to catch it.
Across the chain: Science
Contents
In brief
- A quaker – a bean that stays pale after roasting – comes from immature green coffee, so it is not a roast defect; roast defects come from the roast itself.8,9
- In a study of one coffee roasted to a normal and five defect profiles, consumers liked the normal roast best, significantly more than every defect roast except the baked one.1
- Baked and underdeveloped roasts showed no clear visible sign: their colour was near normal in the defect study, and a baked roast matched the others in a second study.1,3
- Practitioners name four surface marks – scorching, tipping, facing and craters – and tie the first three to too much surface heat; no peer-reviewed study of them was found.5,12,13,14
- In the same defect study, dark and scorched roasts were rated most intense, burnt and bitter; a roast stopped 10 s after first crack was thinnest and least liked.1
What counts as a roast defect#
A roast defect is a flaw that comes from the roasting process. It does not come from how the coffee was grown, processed or stored as green coffee.1 The main sensory study on the subject focused on defects of the roasting process. It left flaws from production, processing and storage of the green coffee for future work.1
In that study, six experts from the Specialty Coffee Association of Europe (SCAE) designed one normal roast and five defect roasts.1 All six roasts used the same washed Kenyan arabica coffee.1 The defects – light, scorched, dark, baked and underdeveloped – were meant to represent faults commonly found on the market.1 The normal roast had average levels of the key aroma compounds and a balanced sensory profile.1 Consumers liked it best, and it was significantly preferred over every defect roast except the baked one.1
Defects fall into two groups. Some come from the time–temperature path of the whole batch: underdeveloped, baked and over-roasted. These show mainly in the cup and in the roast log, the record of temperature and time for each batch. Others are marks on individual beans that roasters see when they inspect a batch: scorching, tipping, facing and craters. Uneven roasting can be either. The second group is described almost only by practitioners; no peer-reviewed study of these marks was found (see Open questions).12,13,14,15
A · Marks on the bean surface
- Scorching. Practitioners use the name for small brownish-black or black specks on the surface, which they put down to overly hot metal.12 According to a roasting textbook chapter, a drum turning too fast can press beans against its wall and, at worst, scorch the flat sides.5
- Tipping. Practitioners use the name for ends of the bean that look burnt.13,15 No published study of it was found (see Open questions).
- Facing. In roasting practice, facing is the name for scorch marks on the flat side, or face, of the bean.14 One practitioner uses the name for scorching late in the roast.12
- Crater. Practitioners use the name for a small round or oblong piece blown out of the bean during roasting.12,15 They see craters in dark roasts at or past second crack, and know of no researched cause.12,15
B · Colour, outside and inside
- Reference. A normally roasted bean. In the defect study, the normal roast measured 74.4 on the Agtron colour scale, where lower numbers mean darker roasts.1
- Over-roasted. Much darker (Agtron 45.7 for the dark roast of the study).1 According to a roasting textbook chapter, in extremely dark roasts oil can reach the surface during the last stages of roasting.5
- Uneven inside the bean. According to a roasting textbook chapter and a doctoral thesis, in extremely fast roasts the surface can be over-roasted while the core stays under-roasted.5,7
- Quaker. A pale bean among normally roasted ones – caused by unripe cherries, not by the roast.8,9
Profile defects: underdeveloped, baked, over-roasted#
Underdeveloped
What it is. The word is used in two ways. A practitioner paper notes that under-development has been blamed on too short a development time and on too low an end temperature.12 Development time is the time from first crack, when the beans start to crack audibly, to the end of the roast. The paper keeps “under-developed” for the first and calls the second “under-roasted”.12
In the defect study, the “Light” roast came closest to this idea: it followed the normal curve but was stopped 10 s after first crack.1 The study’s “Underdeveloped” sample was something else. It was charged (put into the roaster) into cooler air, and its temperature curve stalled at the start.1
How it looks. The short-developed Light roast ended much lighter than the normal roast; the slow-start roast ended at almost the same colour.1
In the cup. Trained assessors rated the Light roast lower than all other samples in most attributes, and higher in citrus, sweet and roasted rye bread.1 Consumers called it thin, mild and bland, linked it with hay or straw, and liked it least of the six.1 The slow-start roast was expected to give a flat, slightly sour cup, but its aroma composition was close to the normal roast.1
Short development is not a defect in itself. In a 2020 study that varied only development time at one colour, the shortest development (90 s) scored highest for acidity, fruitiness and clean cup.3 The defect lies at the extreme, as in the 10 s Light roast, not in short development as such.1,3
Check and avoid. Roast colour and the time from first crack both reveal the Light roast: it differed from the normal roast in both.1 The slow-start roast differed in neither colour nor development time.1 It differed only in its start and total time, so it is found in the roast log and the cup.1
Baked
What it is. A 2020 development-time study defines a baked roast by a drastically stretched time after first crack, with little or no temperature rise.3 A multi-study paper reports that the Specialty Coffee Association (SCA) roaster certification includes it as a defect of overly extended development time.4 The word “stalled” is used for both underdeveloped and baked roasts, at different moments: the slow-start roast of the defect study stalled at the beginning.1 A baked roast stalls after first crack.3
How it looks. There is no distinct sign. In the defect study, the baked roast ended close to the normal roast’s colour.1 In the development-time study, all roasts – baked included – were brought to the same colour.3
In the cup. In the development-time study, the baked roast had the lowest sweetness and acidity and the highest bitterness, roasted and nutty–chocolate notes.3 The authors note that it gained characteristics akin to darker roasts.3
Check and avoid. Colour did not reveal the baked roasts in either study.1,3 The roast log does: a long development with little temperature rise after first crack is the definition of the defect.3 In the cup, according to the multi-study paper, the SCA roaster certification characterises it by strong roasted-bread notes.4
Over-roasted
What it is. A roast taken too far. In the defect study, the dark roast followed the normal curve but ran longer.1 According to a roasting textbook chapter, the reactions late in the roast turn exothermic – they release heat of their own – and speed up.5 The chapter notes that a few seconds can separate the desired roast degree from an over-roasted product.5 It adds that an uncontrolled roast can catch fire.5
How it looks. The dark roast of the defect study was much darker than the normal roast.1 According to the textbook chapter, oil moves to the surface faster in dark roasts.5 In extremely dark roasts, oil can appear during the last stages of roasting.5
In the cup. Trained assessors rated the dark roast, together with the scorched one, highest for burnt, bitter, tobacco and liquorice, and gave it the weakest acidic aftertaste.1 Consumers rated it the strongest and most intense, and called it unpleasant more often than any other sample.1 A chemistry chapter describes over-roasting as giving unbalanced, burnt, harsh off-notes.6 According to the roasting textbook chapter, over-roasted coffee may show a strong “roasty” note but often a thin mouthfeel.5
Check and avoid. The checks are a colour measurement and a look for surface oil.1,5 According to the textbook chapter, the roast should be stopped abruptly at the desired degree, with efficient pre-cooling or cooling.5
Marks on the bean: scorching, tipping, facing, craters#
These four are practitioners’ names. The sources that describe them are a roaster manufacturer’s technical paper, a retailer’s glossary and a trade magazine.12,13,14,15 They do not agree on every definition (see Contested).12,13,14,15 A roasting textbook chapter supports the mechanism behind one of them.5 In its account, a drum that turns too fast lets the beans “hug” its hot wall longer.5 This over-develops the outside of the bean and, at worst, scorches the flat sides.5 A doctoral thesis on roasting describes a roaster’s main task as mixing the beans well enough for a homogeneous roast and preventing scorching.7
Scorching
Look and cause. A roaster manufacturer’s paper describes small brownish-black or black specks on the surface.12 It sees them usually before the bean turns brown, and puts them down to contact with metal that is too hot.12 Another practitioner source describes a larger dark spot, typically on the flat side, from a surface too hot for the bean.15
In the cup. The defect study’s “Scorched” sample was defined by its roast profile, not by marks on the beans.1 It was charged into hotter air and finished sooner than the normal roast.1 Trained assessors rated it as intense, burnt and bitter as the dark roast, and found little difference between the two.1 Yet the aroma compositions of the two roasts differed strongly.1 Consumers rated it less intense than the dark roast.1
Prevention. The doctoral thesis names good mixing.7 The textbook chapter names a drum speed that does not press beans against the wall.5 The manufacturer’s paper names charging too hot, too little airflow and too much burner early in the roast as causes.12 It also argues that a drum not heated directly is hard to overheat – a claim about its own design.12
Tipping
Look and cause. In roasting practice, tipping is the name for ends of the elongated bean that look burnt.13 One practitioner sees this charred spot most often early in the roast, before browning, and puts it down mainly to roasting too rapidly.12 A retailer’s glossary names an overheated roaster at the start, too much coffee in the drum, or possibly too little air movement.13
In the cup. A retailer’s glossary describes burnt or smoky flavours and a lack of sweetness.13 A trade-magazine article says a tipped roast can still give a good cup, and that the cup should decide.15 No sensory study opened for this entry has isolated tipping.
Facing
Look and cause. In roasting practice, facing is the name for scorch marks on the flat side, the “face”, of the bean.14 A practitioner paper reserves the name for scorching late in the roast.12 It gives two causes: the swollen batch forces the beans against the drum wall, or the drum metal overheats towards the end.12 The textbook’s drum-speed mechanism fits this pattern.5
In the cup. Not established; practitioners treat facing as a sign of scorching.14
Craters
Look and cause. In roasting practice, a crater is a small round or oblong piece blown out of the bean during roasting, sometimes called chipping.12 One practitioner sees craters only in dark roasts, at or past the start of second crack.12 The same practitioner knows of no researched reason and suggests pressure trapped in the brittle bean.12 A trade-magazine article places craters near or into second crack, when built-up gas is released so fast that it blows a piece off.15
In the cup. Practitioners disagree. One practitioner paper says craters do “not seem to have any perceived impact on flavor”.12 A trade-magazine article treats them as a possible cause of unwanted flavours.15
Uneven roasts#
Within one bean. According to a roasting textbook chapter, rapid heating can create a steep temperature gradient from surface to core.5 In extreme cases the outside is over-roasted while the core stays under-roasted, which can give “burnt” and “greenish” notes in the same cup.5 A doctoral thesis adds that colour then becomes a less reliable guide, because the interior is less roasted than the outside.7
Between beans. According to the textbook chapter, when a blend is roasted in one batch, its components can end with visibly different roast degrees.5 The chapter describes split roasting – roasting the components separately and blending afterwards – as letting each be roasted to its own optimum.5 It notes that some roasters accept or even want the variation from a green blend, and that the decision rests on the cup.5
Check and avoid. The sources name three measures: looking at the whole beans for colour differences5, mixing the batch well7 and avoiding extreme heating rates.5
Quakers: not a roast defect#
According to a cupping and grading chapter, a quaker is a bean that stays light, pale brown after roasting.8 It looks distinctly different from the fully roasted beans around it.8 The chapter traces it to a cherry picked before it was fully ripe, so the seed did not develop fully.8 A 2021 study confirmed that the colour and sensory characteristics of quakers follow from the chemical composition of the immature seed.9 It found that their aroma compounds came from precursors already present in the immature green coffee.9 According to the chapter, the remedy lies before roasting: high-quality coffees are hand-picked by pickers trained to take only ripe, red cherries.8
Reference table#
Defect, visible sign, likely cause, cup and check, as reported in the sources. “Not established” means no source opened for this entry supports a statement. Cells marked with a practitioner source (12–15) rest on Tier C sources and are practitioner knowledge, not tested findings.
| Defect | Visible sign | Likely cause in the roaster | Cup | Check |
|---|---|---|---|---|
| Underdeveloped (short development) | Much lighter colour: Agtron 116.6 vs 74.4 in one study1 | Roast ended 10 s after first crack (8:40 vs 11:25 min)1 | Thin, mild, bland, hay/straw; least liked1 | Colour; development time in the log1 |
| Underdeveloped (slow start) | None: Agtron 74.9 vs 74.41 | Charged into 135 °C air; curve stalled at the start1 | Close to normal; “caramel” ticked less often1 | Start of the roast log; cupping1 |
| Baked | None: colour close to normal1,3 | Development stretched with little or no temperature rise3,4 | Less sweet and acidic, more bitter and roasted; roasted bread3,4 | Development time and temperature rise after first crack3 |
| Over-roasted | Dark (Agtron 45.7)1; surface oil in extremely dark roasts5 | Roast continued too long1; late exothermic phase not stopped in time5 | Burnt, bitter, tobacco, liquorice; intense1; harsh6 | Colour; surface oil1,5 |
| Scorching (surface) | Dark specks or spots, often on the flat side12,15 | Contact with too-hot metal12,15; drum too fast5; poor mixing7 | Not established for the marks; the high-heat “scorched” profile was rated burnt and bitter1 | Inspect the beans12 |
| Tipping | Burnt ends of the bean13,15 | Practitioners: roasting too rapidly, overheated start, overfilled drum, low airflow12,13 | Practitioners: burnt or smoky, less sweet13 | Inspect the beans13 |
| Facing | Scorch marks on the flat side14 | Beans pressed against the drum wall late in the roast; overheated drum5,12 | Not established | Inspect the beans14 |
| Craters | Small round or oblong piece blown out12 | Not established; pressure inside the bean suggested12,15 | Not established (practitioners disagree)12,15 | Inspect the beans12 |
| Uneven (within the bean) | Colour less reliable7 | Extremely fast heating; steep surface-to-core gradient5 | Burnt and greenish at once5 | Not established |
| Uneven (between beans) | Visibly different roast degrees in the batch5 | Blend components roasted together; poor mixing5,7 | Not established | Inspect the beans5 |
| Quaker (not a roast defect) | Pale, light-brown beans8 | None – immature green coffee8,9 | Quakers of Agtron 82.8 or lighter lowered sensory quality of natural specialty coffees only from 7 in a cup of 65 beans9 | Inspect the roasted sample8 |
Profile defects in detail#
Roast times, charge temperatures, colours and aroma markers of the defect study, and the detail behind the baked defect.
Underdeveloped
The Light roast of the defect study followed the normal curve but ended at 8:40 min instead of 11:25 min.1 The study’s Underdeveloped sample was charged into air at 135 °C instead of 210 °C and roasted for 20:20 min.1 The Light roast ended at Agtron 116.6 against 74.4 for the normal roast; the slow-start roast ended at 74.9.1
The proposed aroma marker of the Light roast is indole, a compound described as flowery and mothball-like.1,2 Consumers ticked “caramel” for the slow-start roast less often than for the normal roast.1 Its marker is 2,5-dimethylfuran, with an ether-like odour.1,2
Baked
In the defect study the baked roast ended at Agtron 68.3, close to the normal 74.4.1 In the development-time study, the common colour of all roasts was Agtron 76.3 To get there, the baked roast was ended at 191.1 °C after 390 s of development, the fastest at 204.1 °C after 90 s.3
In the development-time study, the slow roast was close to the baked one on most attributes.3 A multi-study analysis tied longer development time – not a longer time to first crack – to more “roasted bread” flavour, which the SCA certification uses to characterise the baked defect.4 In the defect study, trained assessors found the baked roast close to the normal one, but with a weaker acidic aftertaste; consumers ticked “sweet” for it far less often and liked it almost as much as the normal roast.1 Maltol (caramel-like) was proposed as its aroma marker.1,2 The development-time study found pyridine rising with development and described it as a previously proposed marker of the baked defect3; the aroma study it cites names pyridine among the compounds that rose in the baked roast but chose maltol as the marker.1,2
Over-roasted
The dark roast of the defect study ran 13:45 min against 11:25 min for the normal roast.1 It ended at Agtron 45.7 against 74.4 for the normal roast.1 Phenol was proposed as its aroma marker.1,2
Scorched profile
The defect study’s Scorched sample was charged into air at 275 °C and finished in 7:40 min against 11:25 min for the normal roast.1 Its proposed aroma marker is 4-ethyl-2-methoxyphenol.1,2
The defect study in numbers#
The 2019 defect study remains the only published sensory study of intentionally produced roast defects found for this entry. The roasts were made on a 1 kg drum roaster.1 The coffee was roasted twice, once for the trained panel and once for the consumer test and aroma analysis.1 Nine trained assessors rated 18 attributes on a 15 cm line scale, three times per sample; 83 regular coffee drinkers rated liking on a 9-point hedonic scale and ticked words from a list of 30 (check-all-that-apply).1 The brews were French press, served at 60 °C.1
| Profile | Charge air temperature | Development time | Total roast time | Agtron | Mean liking (SD) |
|---|---|---|---|---|---|
| Normal | 210 °C | 2:40 min | 11:25 min | 74.4 | 6.0 (1.8) a |
| Light | 210 °C | 0:10 min | 8:40 min | 116.6 | 4.2 (2.1) d |
| Scorched | 275 °C | 1:50 min | 7:40 min | 66.0 | 5.6 (2.2) bc |
| Dark | 220 °C | 4:45 min | 13:45 min | 45.7 | 5.1 (2.2) c |
| Baked | 230 °C | 6:20 min | 18:00 min | 68.3 | 5.7 (1.9) ab |
| Underdeveloped | 135 °C | 2:30 min | 20:20 min | 74.9 | 5.4 (2.0) bc |
Source: Giacalone et al., Tables 1 and 5.1 Means sharing a letter do not differ significantly (Tukey, p < 0.05).1
Data table
| mean liking, 9-point scale | Source | |
|---|---|---|
| Normal (reference) | 6.0 (SD 1.8) a | 1 |
| Baked | 5.7 (SD 1.9) ab | 1 |
| Scorched | 5.6 (SD 2.2) bc | 1 |
| Underdeveloped | 5.4 (SD 2.0) bc | 1 |
| Dark | 5.1 (SD 2.2) c | 1 |
| Light | 4.2 (SD 2.1) d | 1 |
Data table
| Agtron (spectrophotometric roast colour) | Source | |
|---|---|---|
| Light | 116.6 · dev. 0:10 | 1 |
| Underdeveloped | 74.9 · dev. 2:30 | 1 |
| Normal (reference) | 74.4 · dev. 2:40 | 1 |
| Baked | 68.3 · dev. 6:20 | 1 |
| Scorched | 66.0 · dev. 1:50 | 1 |
| Dark | 45.7 · dev. 4:45 | 1 |
Selected mean ratings of the trained panel (15 cm scale) show how the defects separate:1
| Attribute | Normal | Baked | Dark | Light | Scorched | Underdev. |
|---|---|---|---|---|---|---|
| Intensity | 9.2 b | 9.3 b | 11.5 a | 3.8 c | 11.9 a | 8.2 b |
| Burnt | 9.1 b | 8.9 b | 12.0 a | 3.9 c | 11.8 a | 8.0 b |
| Bitter | 9.6 b | 10.3 b | 12.2 a | 4.8 c | 12.4 a | 9.1 b |
| Citrus | 5.7 b | 4.4 bc | 1.9 c | 8.8 a | 3.0 c | 5.2 bc |
| Sweet | 3.8 b | 3.0 bc | 2.3 c | 5.0 a | 2.2 c | 3.7 b |
| Acidic (aftertaste) | 10.3 a | 7.9 b | 5.7 c | 8.6 ab | 8.2 b | 9.3 ab |
| Complexity | 8.8 a | 8.4 ab | 6.5 b | 5.2 c | 7.3 ab | 7.6 ab |
Source: Giacalone et al., Table 4.1 Means in a row sharing a letter do not differ significantly (Tukey, p < 0.05).
Three patterns stand out. The light roast sits at one end and the dark and scorched roasts at the other; the normal, baked and underdeveloped roasts sit near the middle and were poorly separated by the first dimension of the analysis.1 The normal roast scored highest in complexity: low intensity (light) and high intensity (dark) both lowered it, an inverse U-shape the authors supported with a quadratic model (p = 0.06).1 And the normal roast was rated high in acidity, which the authors link to the Kenyan coffee chosen; in another coffee, high acidity could itself be seen as a defect.1
Aroma markers#
A 2016 aroma study analysed the same six roasts by headspace solid-phase micro-extraction and GC–MS. Of 37 known potent coffee odorants, some changed significantly with each defect, and one marker was proposed for each: indole (light), 4-ethyl-2-methoxyphenol (scorched), phenol (dark), maltol (baked) and 2,5-dimethylfuran (underdeveloped).2 The authors suggest such markers could become screening tools for roast defects in production.2
The markers follow the chemistry of the profiles. The scorched roast showed higher levels of 4-ethyl-2-methoxyphenol, pyridine, phenol and difurfuryl ether; the dark roast more phenolic compounds; the baked roast a slight overall increase, largest for maltol, difurfuryl ether and pyridine.1 Most aroma compounds – pyrazines, aldehydes, alcohols, sulphides, pyrroles and furans – rose with longer and hotter roasting, and burnt and astringent ratings rose with them.1 In the development-time study, longer development degraded the acids and chlorogenic acids, raised pyridine and some alcohols with fusel or roasty character, and slightly raised Maillard-derived pyrazines.3 All acids were degraded to the same extent, so their ratio did not change.3
Aroma data and perception do not always match. The underdeveloped roast had the most acetic acid of the six (41.64 ppm, against 31.04 ppm for the normal and 20.72 ppm for the light roast), yet it was not perceived as more acidic.1 And the panel found the dark and scorched roasts nearly identical, although their aroma compositions were clearly different.1
Development time and the baked defect#
| Development time study (one Colombian coffee, Agtron 76) | Fast | Medium | Slow | Baked |
|---|---|---|---|---|
| Development time | 90 s | 143 s | 266 s | 390 s |
| End temperature | 204.1 °C | 201.2 °C | 198.4 °C | 191.1 °C |
| Sweetness | 8.7 | 8.3 | 7.2 | 6.3 |
| Acidity | 8.5 | 7.6 | 5.8 | 5.4 |
| Bitter | 5.7 | 6.5 | 8.0 | 8.6 |
| Roasted | 5.4 | 6.3 | 8.3 | 8.8 |
| Clean cup | 7.5 | 6.7 | 5.5 | 5.5 |
Source: Alstrup et al.; first crack at about 570 s; 46 experienced tasters; the paper does not state where the temperature was measured.3 Slow and Baked did not differ significantly in acidity, bitterness, roasted notes or clean cup.3
The table shows why a baked roast is hard to see: all four roasts had the same colour.3 The baked roast reached it at a lower end temperature, after a long, flat development.3 The step from slow to baked was small on the panel’s scales; the step from fast to slow was large.3
Heat at the bean surface#
According to a roasting textbook chapter, high rates of heat transfer at the bean surface can create a substantial temperature gradient from surface to core; in extreme cases of fast roasting, the outer zone is over-roasted while the core stays under-roasted.5 The chapter adds that a drum brings in conduction: beans lying against a hot wall for too long can over-develop on the outside and scorch on the flat side.5 This is the physical basis practitioners give for scorching and facing.5,12
The heat-transfer literature does not reach the scale of tips and edges. A 2017 single-bean model assumes a spherical bean with spherical symmetry10; a 2011 model used both a simplified semi-elliptical shape and a 3D scan of a real C. arabica bean, and compared its predictions with thermocouples inserted at the centre of the bean.11 No model or measurement opened for this entry describes how hot the tips, edges or flat face get compared with the rest of the bean.
Open questions#
- Surface marks. Tipping, facing and craters have no peer-reviewed description, cause or sensory study among the sources opened; practitioners disagree on causes and on whether craters affect flavour.12,13,14,15 A controlled study that creates each mark on purpose and tests its effect in the cup is missing.
- One coffee, one market. The main defect study used one Kenyan arabica coffee and defect definitions rooted in Northern European roasting practice; the authors caution that results may not hold for other coffees and regions.1
- Consumer relevance. The liking range was narrow (4.2 to 6.0), and the lack of a significant difference between the normal and baked roasts may be due to limited statistical power.1 Liking for the dark and light roasts was bimodal: a sizeable minority liked them most.1
- Names. “Underdeveloped”, “baked” and “scorched” mean different things in the study and in practice (see Contested).1,3,12
Key studies#
Giacalone et al., 2019 – Common roasting defects. One washed Kenyan arabica coffee was roasted to a normal profile and five defect profiles designed by SCAE experts, then studied by GC–MS, a trained panel and 83 consumers. Light sat at one end of the sensory space, dark and scorched at the other; the normal roast was balanced, most complex and best liked, though not significantly ahead of the baked roast.
Yang et al., 2016 – Aroma markers of roast defects. The same six roasts analysed for 37 potent coffee odorants by headspace GC–MS. Each defect changed the aroma profile in its own way, and one marker compound was proposed per defect as a basis for industrial screening.
Alstrup et al., 2020 – Development time up to baked. Four roasts of one coffee at the same colour, with development from 90 to 390 s, the longest defined as baked. Longer development lowered sweetness, acidity and clean cup and raised bitterness and roasted notes, while NMR and GC–MS tracked the loss of acids and the rise of pyridine.
Münchow et al., 2020 – Which roast phase drives flavour. Pools eight sensory studies, the first of them the defect study. Across seven studies at constant colour, development time, not time to first crack, was tied to more "roasted bread" – the note that characterises the baked defect.
Rabelo et al., 2021 – Quakers in natural coffees. Quaker beans of three colours were added in seven amounts to natural specialty coffees. Quakers of Agtron 82.8 or lighter hurt the cup only from 7 quakers in a cup of 65 beans, and their aroma compounds came from precursors in the immature green coffee – a green-coffee flaw, not a roast flaw.
- Meaning of "underdeveloped": a practitioner paper notes that under-development has been put down both to too short a development time and to too low an end temperature, and keeps the word for the first12; in the main sensory study, the roast with a 10 s development was called "Light", and "Underdeveloped" was a roast charged into air at 135 °C whose temperature curve stalled at the start.1 The aroma study on the same samples uses the same labels.2
- Definition of "baked": a development-time study defines it as time after first crack extended drastically with little to no increase in temperature3; a multi-study paper describes the defect in the Specialty Coffee Association (SCA) roaster certification as an overly extended development time4; the baked sample of the defect study was charged into hotter air (230 °C) and roasted for 18:00 min against 11:25 min for the normal roast1; a practitioner paper reports that roasters disagree on the cause.12
- Scorching, facing and tipping overlap in practitioner use: one glossary calls facing and tipping the telltale signs of scorching14; a practitioner paper calls scorching that happens late in the roast "facing"12; a trade-magazine article separates tipping (the tip) from scorching (typically the flat side).15
- Causes of tipping: "roasting the coffee too rapidly"12; an overheated roaster at the start, an overfilled drum or too little air13; or any temperature too high for the bean to take up the heat.15 None of these has been tested in a published study.
- Craters: a practitioner paper finds no researched reason and says they do not seem to affect flavour12; a trade-magazine article places them near or into second crack, when built-up gas is released too fast, and treats them as a possible cause of unwanted flavours.15
- Fast roasts: a 2020 development-time study notes that extremely fast roasts are traditionally considered defects because of a theoretical gradient of roast degree from surface to centre3, and a roasting textbook chapter describes surface over-roasting with an under-roasted core in extreme cases5; yet in a development-time study at one colour the fastest development gave the highest acidity, fruitiness and clean-cup scores.3
Coffearium (2026). Roast defects. Coffearium – The open coffee compendium. Retrieved 22 October 2026, from https://coffearium.org/entries/roast-defects/
@misc{coffearium2026-roast-defects,
author = {{Coffearium}},
title = {{Roast defects}},
year = {2026},
howpublished = {Coffearium -- The open coffee compendium},
url = {https://coffearium.org/entries/roast-defects/},
urldate = {2026-10-22},
note = {Retrieved 22 October 2026}
}Entries are updated when the evidence changes; the retrieval date tells your readers which version you used.
Sources
- 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
- AYang, N.; Liu, C.; Liu, X.; Degn, T. K.; Munchow, M.; Fisk, I. D. (2016). Determination of volatile marker compounds of common coffee roast defects. Food Chemistry 211, 206–214. doi:10.1016/j.foodchem.2016.04.124 · other copies
- 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
- 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
- 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
- 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
- BSchenker, S. (2000). Investigations on the Hot Air Roasting of Coffee Beans (Diss. ETH No. 13620). ETH Zürich (doctoral thesis).
- BLingle, T. R.; Menon, S. N. (2017). Cupping and Grading – Discovering Character and Quality. In: Folmer, B. (ed.), The Craft and Science of Coffee, Academic Press, ch. 8, pp. 181–203. doi:10.1016/B978-0-12-803520-7.00008-6 · other copies
- ARabelo, M. H. S.; Borém, F. M.; Lima, R. R.; Alves, A. P. C.; Pinheiro, A. C. M.; Ribeiro, D. E.; Santos, C. M.; Pereira, R. G. F. A. (2021). Impacts of quaker beans over sensory characteristics and volatile composition of specialty natural coffees. Food Chemistry 342, 128304. doi:10.1016/j.foodchem.2020.128304 · other copies
- 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
- AFabbri, A.; Cevoli, C.; Alessandrini, L.; Romani, S. (2011). Numerical modeling of heat and mass transfer during coffee roasting process. Journal of Food Engineering 105(2), 264–269. doi:10.1016/j.jfoodeng.2011.02.030 · other copies
- CHoos, R. (2020). How to Avoid Roasting Defects (technical paper). Loring Smart Roast (roaster manufacturer).
- CSweet Maria's Tipping (Coffee Library glossary). Sweet Maria's Coffee Library (green-coffee retailer).
- CSweet Maria's Facing (Coffee Library glossary). Sweet Maria's Coffee Library (green-coffee retailer).
- CNeil (byline only) (2021). Roaster Focus: What is happening to coffee beans inside the roaster?!. Coffee Magazine (True North Media); first posted by a roaster manufacturer.
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”.
Discussion
A space for readers to share experience from the farm, the roastery or the bar, ask each other and offer other views. The Coffearium team does not reply here. Please keep to the community rules; to change the entry itself, use Suggest a change.