Topic hub — 2026-10-22
Does the way coffee is brewed change its bitterness?
This page collects studies that compared brewing methods for the same beans, with people rating bitterness. It is limited to comparisons where brewing alone can be examined, not roast or species (roast is covered on the dark-roast page and species on the arabica and robusta page). In the included studies, hot brews were more bitter than cold brews in most results, but some pairs of the same beans showed no significant difference, and in many studies the values are only in figures that we could not obtain. Studies that varied water temperature within the hot range detected no significant difference in bitterness; equipment comparisons also differ in grind, in the amounts of coffee and water and in extraction style; and for water, depending on what was added and how much, some conditions showed a significant difference and others did not. Beans, strength, serving temperature, scales and raters differ by study, so it cannot be said across studies how much more bitter coffee is at what temperature. Ultra-high-pressure extraction is an industrial process, not home brewing; in the list its labels begin with "Ultra-high pressure" and its rows are placed at the end. Bitterness is read apart from astringency and sourness. This page lines up the included studies together with their conditions: the beans, the roast, how they were brewed and served, and who rated them on what scale. It then separates what was established within one experiment, what differences might explain, and what this comparison alone cannot separate. It is about perceived taste, not a recommendation on how to brew. Coffee bitterness as a whole is covered on the coffee bitterness page.
On this page (10 sections)
- What this page establishes
- What the included studies allow us to say
- Bitterness ratings at a glance
- Conditions and results, study by study
- What was established within an experiment
- Candidate explanations for the differences
- What is not known — what this comparison alone cannot separate
- How studies and rows were counted
- What we looked at
- Read in the encyclopedia
What this page establishes
- That hot brews of the same beans were more bitter than cold brews in most results, while some pairs of the same beans showed no significant difference, and that studies differ in how they matched strength and serving temperature.
- That studies varying water temperature within the hot range detected no significant difference in bitterness, which does not show that bitterness was the same.
- That equipment comparisons also differ in grind, in the amounts of coffee and water and in extraction style, so the effect of the device itself cannot be separated, and that no result of a bitterness comparison between espresso and filter is reported.
- That for water, depending on what was added and how much, some conditions showed a significant difference in bitterness and others did not (the specific conditions are in the sentences below), and that a comparison of commercial waters cannot be called an effect of hardness alone.
- That ultra-high-pressure extraction is not home brewing, and while some of its conditions were significantly less bitter than conventional cold brew, pressure and time differ together.
What the included studies allow us to say
- In studies where people compared the bitterness of the same beans brewed with hot water and as cold brew (steeped for a long time in cold or room-temperature water), the hot brew was more bitter in most results. In a study that diluted both to 2% strength and served them at the same temperature (Cai et al. 2022), bitterness was 4.14 hot and 2.43 cold (1–5 points), significantly higher hot. A study serving a hot drip and a 4 °C immersion of the same beans at 7 °C (Lopane et al. 2024), a comparison of immersion temperatures pooled over 9 coffees (Batali et al. 2022) and a study comparing immersion temperatures of the same beans by roast (Liang et al. 2024) also say in the text that the hot brew was significantly more bitter (values are in figures and unconfirmed). On the other hand, in a study where 120 consumers rated the same beans brewed in a coffee maker and as 4 °C cold brew (Heo et al. 2019), no significant difference was detected in 2 of the 4 pairs. [rows: Cai 2022 Hot brew (coffee machine) vs cold brew (8 h at room temperature); Lopane 2024 Hot drip vs 4 °C immersion (storage day 0); Batali 2022 Overall result (pooled over origins and roasts): 92 °C vs 4 °C immersion (9 coffees); Liang 2024 92 °C and 22 °C vs 4 °C immersion (darker roast); Liang 2024 92 °C vs 4 °C and 22 °C immersion (lighter roast); Heo 2019 Coffee maker vs cold brew (blend); Heo 2019 Coffee maker vs cold brew (Colombia); Heo 2019 Coffee maker vs cold brew (Folgers); Heo 2019 Coffee maker vs cold brew (Kenya)]
- Studies that varied water temperature within the hot range detected no significant difference in bitterness (Batali et al. 2020 at 87, 90 and 93 °C with strength and extraction matched; Izumi and Takaya 2008 at 85, 90 and 95 °C). In the same Batali et al. 2020, the higher the strength (TDS), the more bitter. For equipment, the siphon scored more bitter than paper drip (about 3.3 and about 2.3), but grind and amounts also differed; espresso and filter had medians of 4.0 and 1.2, but bitterness was not among the significant attributes. For water, brewing with 300 mg/L bicarbonate added was significantly more bitter than ultrapure water (9.5 and 6.9), while water with only 264 mg/L calcium added showed no significant difference (Rune et al. 2026). [rows: Batali 2020 Drip at 93 °C vs 87 °C (and 90 °C), strength and extraction matched; Batali 2020 Strength (TDS) about 1.5% vs about 1% (pooled over temperature and extraction); Izumi and Takaya 2008 Paper drip at 95 °C vs 85 °C (and 90 °C) (3 coffees); Izumi and Takaya 2008 Siphon vs paper drip (coffee A; grind also differs); Vezzulli 2022 Espresso vs filter (arabica); Rune 2026 Water with 300 mg/L bicarbonate added vs ultrapure water; Rune 2026 Water with 264 mg/L calcium added vs ultrapure water]
- Beans, strength, serving temperature, scales and raters differ by study, so it cannot be said across studies how much more bitter coffee is at what temperature. Ultra-high-pressure extraction (100–500 MPa) is an industrial process, not home brewing; in the list its labels begin with "Ultra-high pressure" and its rows are placed at the end. Bitterness is read apart from astringency and sourness. This page lines up studies of perceived taste and is not a recommendation on how to brew. [rows: Chen 2023 Ultra-high pressure: 500 MPa, 20 min vs conventional cold brew (atmospheric, 5 °C, 12 h); Chen 2023 Ultra-high pressure: 100 MPa vs 500 MPa (both 20 min)]
Bitterness ratings at a glance
Human bitterness ratings only. Beans, brewing and scales differ by study (point scores, line scales, medians on short scales, consumer category scales), so values are not compared across studies. The "Direction" column shows which sample was more bitter ("... more bitter"); in every row on this page the larger value is the more bitter one. The direction shows which published mean or median is larger, or what the text states, and is not necessarily the result of an individual test between the two conditions shown. Whether the difference is significant is shown by the test in the same "Direction / test" column. Rows where no significant difference was detected and rows whose values are only in figures we did not obtain are kept as they are. Labels begin with a tag for rows of a different nature: "Ultra-high pressure" marks an industrial process, not home brewing, placed at the end of the list; "Overall result (pooled over origins and roasts)" marks rows that are not per-coffee comparisons; and "Reference (serving temperature and session differ)" marks rows that cannot be read as comparisons of brewing alone, marked as not comparable. Select a study name to jump to its detailed table.
Results in which people rated bitterness only. Where each value comes from (a table or a reading from a figure) is shown in its row; scales and methods differ between studies, so compare within a row, not across rows.
| Study | Comparison and condition | Measured as | Brewing compared | Comparison | Direction / test |
|---|---|---|---|---|---|
| Cai 2022 | Hot brew (coffee machine) vs cold brew (8 h at room temperature) Same batch of a medium-dark arabica blend. Hot: 60 g grounds to 600 g water, outlet 80 ± 3 °C, about 6 min. Cold: same 60 g and 600 g, 20 ± 3 °C for 8 h, then rinsed with another 120 g water and combined. Both diluted to 2% strength (TDS) and served at 20 ± 3 °C | 1–5 points per the methods (mean ± SD of 7 trained panellists); the sweetness value of 0.71 in the same table falls outside 1–5, so the lower end of the scale is inconsistent within the source | 4.14 ± 0.69 | 2.43 ± 0.53 | Brewing compared more bitter Tested, significant |
| Lopane 2024 | Hot drip vs 4 °C immersion (storage day 0) The same Brazilian arabica (Acaiá, roast degree Gourmet 72.9), same grind, mineral-standardised water. Hot: drip at about 97 °C, coffee to water 1:18, flash-chilled to 7 °C in an ice bath. Cold: immersion at 4 °C for 4 h, 1:14; there is also a 25 °C immersion under the same conditions. Bottled, stored at 7 °C and served at 7 °C | intensity from 0 (none) to 15 (5 coffee professionals, 3 replicates) | Figure not obtained, so values unconfirmed | Figure not obtained, so values unconfirmed | Brewing compared more bitter Tested, significant |
| Lopane 2024 | Hot drip vs 4 °C immersion (storage day 42) The same Brazilian arabica (Acaiá, roast degree Gourmet 72.9), same grind, mineral-standardised water. Hot: drip at about 97 °C, coffee to water 1:18, flash-chilled to 7 °C in an ice bath. Cold: immersion at 4 °C for 4 h, 1:14; there is also a 25 °C immersion under the same conditions. Bottled, stored at 7 °C and served at 7 °C | intensity from 0 (none) to 15 (5 coffee professionals, 3 replicates) | Figure not obtained, so values unconfirmed | Figure not obtained, so values unconfirmed | Brewing compared more bitter Tested, significant |
| Heo 2019 | Coffee maker vs cold brew (blend) The same beans brewed as cold brew (9 h at 4 °C, kept refrigerated, used within 22 h) and in a coffee maker (cooled 30 min in an ice bath and refrigerated, used within 10 h); 7 g grounds to 100 mL water and medium grind for both; served at about 8 ± 2 °C, 50 mL. Beans: city-roast blend (Brazil, Colombia, Ethiopia, Guatemala) | intensity on a 9-point category scale (mean of 120 untrained consumers); LSD letters in brackets (values sharing a letter do not differ significantly) | 4.9 (d) | 4.9 (d) | About the same Tested, not significant |
| Heo 2019 | Coffee maker vs cold brew (Colombia) The same beans brewed as cold brew (9 h at 4 °C, kept refrigerated, used within 22 h) and in a coffee maker (cooled 30 min in an ice bath and refrigerated, used within 10 h); 7 g grounds to 100 mL water and medium grind for both; served at about 8 ± 2 °C, 50 mL. Beans: Colombia (Nariño), medium roast | intensity on a 9-point category scale (mean of 120 untrained consumers); LSD letters in brackets (values sharing a letter do not differ significantly) | 4.9 (d) | 4.6 (de) | Brewing compared more bitter Tested, not significant |
| Heo 2019 | Coffee maker vs cold brew (Folgers) The same beans brewed as cold brew (9 h at 4 °C, kept refrigerated, used within 22 h) and in a coffee maker (cooled 30 min in an ice bath and refrigerated, used within 10 h); 7 g grounds to 100 mL water and medium grind for both; served at about 8 ± 2 °C, 50 mL. Commercial Folgers classic roast (medium) | intensity on a 9-point category scale (mean of 120 untrained consumers); LSD letters in brackets (values sharing a letter do not differ significantly) | 6.8 (a) | 5.8 (bc) | Brewing compared more bitter Tested, significant |
| Heo 2019 | Coffee maker vs cold brew (Kenya) The same beans brewed as cold brew (9 h at 4 °C, kept refrigerated, used within 22 h) and in a coffee maker (cooled 30 min in an ice bath and refrigerated, used within 10 h); 7 g grounds to 100 mL water and medium grind for both; served at about 8 ± 2 °C, 50 mL. Beans: Kenya AA, medium roast | intensity on a 9-point category scale (mean of 120 untrained consumers); LSD letters in brackets (values sharing a letter do not differ significantly) | 5.0 (d) | 4.4 (ef) | Brewing compared more bitter Tested, significant |
| Batali 2022 | Overall result (pooled over origins and roasts): 92 °C vs 4 °C immersion (9 coffees) 9 coffees (3 arabica origins, each at 3 roast levels) immersed at 100 g grounds to 500 g water at 4 °C (22 °C water poured, then refrigerated 24–36 h), 22 °C (12–14 h) or 92 °C water (left to cool at room temperature, 1–2 h) until extraction stopped; all diluted to 2% strength (TDS) and served at 4 °C; values are the main effect of temperature pooled over origins and roasts | 0–100 line scale (rated at home, 10 panellists analysed, 3 replicates) | Figure not obtained, so values unconfirmed | Figure not obtained, so values unconfirmed | Brewing compared more bitter Tested, significant |
| Batali 2022 | Overall result (pooled over origins and roasts): 22 °C vs 4 °C immersion (9 coffees) 9 coffees (3 arabica origins, each at 3 roast levels) immersed at 100 g grounds to 500 g water at 4 °C (22 °C water poured, then refrigerated 24–36 h), 22 °C (12–14 h) or 92 °C water (left to cool at room temperature, 1–2 h) until extraction stopped; all diluted to 2% strength (TDS) and served at 4 °C; values are the main effect of temperature pooled over origins and roasts | 0–100 line scale (rated at home, 10 panellists analysed, 3 replicates) | Figure not obtained, so values unconfirmed | Figure not obtained, so values unconfirmed | Direction not given Tested, not significant |
| Liang 2024 | 92 °C and 22 °C vs 4 °C immersion (darker roast) The darker roast of the same Central American blend (Agtron 41.8); 100 g grounds to 500 g water, 5 time points over 0–24 h for 4 °C water (refrigerated) and 22 °C water (room temperature) and over 0–60 min for 92 °C water (left at room temperature); concentrate diluted 2-fold (strength not equalised); served at about 6 °C, warming to about 12 °C while tasted; values compare the pooled 5 time points | 0–100 line scale (14 trained panellists, 3 replicates) | Figure not obtained, so values unconfirmed | Figure not obtained, so values unconfirmed | Brewing compared more bitter Tested, significant |
| Liang 2024 | 92 °C vs 4 °C and 22 °C immersion (lighter roast) The lighter roast of the same Central American blend (Agtron 71.8); 100 g grounds to 500 g water, 5 time points over 0–24 h for 4 °C water (refrigerated) and 22 °C water (room temperature) and over 0–60 min for 92 °C water (left at room temperature); concentrate diluted 2-fold (strength not equalised); served at about 6 °C, warming to about 12 °C while tasted; values compare the pooled 5 time points | 0–100 line scale (14 trained panellists, 3 replicates) | Figure not obtained, so values unconfirmed | Figure not obtained, so values unconfirmed | Brewing compared more bitter Tested, significant |
| Liang 2024 | 22 °C vs 4 °C immersion (lighter roast) The lighter roast of the same Central American blend (Agtron 71.8); 100 g grounds to 500 g water, 5 time points over 0–24 h for 4 °C water (refrigerated) and 22 °C water (room temperature) and over 0–60 min for 92 °C water (left at room temperature); concentrate diluted 2-fold (strength not equalised); served at about 6 °C, warming to about 12 °C while tasted; values compare the pooled 5 time points | 0–100 line scale (14 trained panellists, 3 replicates) | Figure not obtained, so values unconfirmed | Figure not obtained, so values unconfirmed | Direction not given Tested, not significant |
| Batali 2020 | Drip at 93 °C vs 87 °C (and 90 °C), strength and extraction matched Honduran arabica roasted on the same day and mixed; commercial drip brewer (flat-bottomed, paper filter); at 87, 90 and 93 °C, 9 targets of strength (TDS) and extraction were reached by adjusting grind, dose and flow, matched across temperatures; served after cooling to 64–66 °C | 15 cm line converted to 0–100 (12 trained panellists, 3 replicates) | Figure not obtained, so values unconfirmed | Figure not obtained, so values unconfirmed | Direction not given Tested, not significant |
| Batali 2020 | Strength (TDS) about 1.5% vs about 1% (pooled over temperature and extraction) Honduran arabica roasted on the same day and mixed; commercial drip brewer (flat-bottomed, paper filter); at 87, 90 and 93 °C, 9 targets of strength (TDS) and extraction were reached by adjusting grind, dose and flow, matched across temperatures; served after cooling to 64–66 °C; the middle level was about 1.25% | 15 cm line converted to 0–100 (12 trained panellists, 3 replicates) | Figure not obtained, so values unconfirmed | Figure not obtained, so values unconfirmed | Brewing compared more bitter Tested, significant |
| Izumi and Takaya 2008 | Paper drip at 95 °C vs 85 °C (and 90 °C) (3 coffees) 3 commercially roasted coffees (A Mocha Sidamo, B Colombia, C Mocha); paper drip, 30 g grounds to 450 mL water, 20 s bloom, 4 min total, distilled water; kept in vacuum flasks so samples were at as similar a temperature as possible when rated | 5-point scoring from 1 to 5 (5 = very strong); 12–14 women | No values given | No values given | Direction not given Tested, not significant |
| Izumi and Takaya 2008 | Siphon vs paper drip (coffee A; grind also differs) Commercially roasted coffee A (Mocha Sidamo, roast degree not given); paper drip (90 °C, 30 g grounds to 450 mL water), electric coffee maker (same amounts) and siphon (20 g grounds to 300 mL water); grind setting 4 (560–450 μm) for the siphon and 9 (710–560 μm) otherwise; the coffee-to-water ratio is 1:15 for both paper drip and siphon; kept in vacuum flasks so samples were at as similar a temperature as possible when rated; the coffee maker was about 2.5 | 5-point scoring from 1 to 5 (5 = very strong); approximate values read by eye from the figure; 12–14 women (Fig. 1) | about 3.3 | about 2.3 | Brewing compared more bitter Tested; significance reported, but which pair differs is unclear |
| Izumi and Takaya 2008 | Commercial hard water vs distilled water (coffee B; a comparison of the commercial waters used) Commercially roasted coffee B (Colombia, roast degree not given); paper drip, 90 °C, 30 g grounds to 450 mL water; waters were distilled, a commercial soft water (hardness 30) and a commercial hard water (Contrex, hardness 1551, unit not given); the soft water was about 3.2 | 5-point scoring from 1 to 5 (5 = very strong); approximate values read by eye from the figure; 12–14 women (Fig. 2) | about 3.7 | about 2.8 | Brewing compared more bitter Tested; significance reported, but which pair differs is unclear |
| Vezzulli 2022 | Espresso vs filter (arabica) 100% Brazilian arabica (same roaster and roasting process, roast degree not given); espresso (92 °C, coffee to beverage 1:2, about 16 g) and filter from a home drip machine (50 g/L); the same beans as moka (76 g/L) 3.4 and Neapolitan pot (72 g/L) 3.1; espresso and filter rated in the same session, moka and Neapolitan in another | median on 0 (absent) to 9 (very intense) (6 trained panellists) | 4.0 | 1.2 | Brewing compared more bitter Test method described; result for this comparison not reported |
| Vezzulli 2022 | Espresso vs filter (robusta) 100% Indian robusta (same roaster and roasting process), brewed as for the arabica; the same beans as moka 3.3 and Neapolitan pot 2.9; arabica versus robusta belongs to another page, and only brewing methods within the same robusta are compared here | median on 0 (absent) to 9 (very intense) (6 trained panellists) | 4.7 | 3.5 | Brewing compared more bitter Test method described; result for this comparison not reported |
| Rune 2026 | Water with 300 mg/L bicarbonate added vs ultrapure water Same batch of Colombian Kachalu (medium roast per the producer); French press, 50 g grounds to 900 ± 5 g water, 92 °C, 4 min; water made by dissolving salts in ultrapure water; served at about 55 ± 2 °C, 30 mL | position on a 15 cm line scale (cm; mean of 10 trained assessors); Tukey letters in brackets (values sharing a letter do not differ significantly) | 9.5 (c) | 6.9 (ab) | Brewing compared more bitter Tested, significant |
| Rune 2026 | Water with 264 mg/L calcium added vs ultrapure water Same batch of Colombian Kachalu (medium roast per the producer); French press, 50 g grounds to 900 ± 5 g water, 92 °C, 4 min; water made by dissolving salts in ultrapure water; served at about 55 ± 2 °C, 30 mL | position on a 15 cm line scale (cm; mean of 10 trained assessors); Tukey letters in brackets (values sharing a letter do not differ significantly) | 6.0 (a) | 6.9 (ab) | Comparison more bitter Tested, not significant |
| Rune 2026 | Water with 264 mg/L calcium and 300 mg/L bicarbonate added vs ultrapure water Same batch of Colombian Kachalu (medium roast per the producer); French press, 50 g grounds to 900 ± 5 g water, 92 °C, 4 min; water made by dissolving salts in ultrapure water; served at about 55 ± 2 °C, 30 mL; water with half of each (132 mg/L calcium and 150 mg/L bicarbonate) was 8.3 (bc), sharing the letter b with ultrapure water | position on a 15 cm line scale (cm; mean of 10 trained assessors); Tukey letters in brackets (values sharing a letter do not differ significantly) | 9.5 (c) | 6.9 (ab) | Brewing compared more bitter Tested, significant |
| Rune 2026 | Water with 132 mg/L calcium and 80 mg/L magnesium added (with or without 60 mg/L bicarbonate) vs ultrapure water Same batch of Colombian Kachalu (medium roast per the producer); French press, 50 g grounds to 900 ± 5 g water, 92 °C, 4 min; water made by dissolving salts in ultrapure water; served at about 55 ± 2 °C, 30 mL; first descriptive analysis (10 trained assessors, 3 replicates) | position on a 15 cm line scale (cm; mean of 10 trained assessors) | 10.0 (no bicarbonate), 9.5 (with bicarbonate) | 9.6 | Depends on the combination Tested, not significant |
| Córdoba 2019 | Coarse vs medium grind (Huila, 14 h cold brew) Colombian Caturra (medium roast); cold brew by immersion at 20 °C, 60 g grounds to 700 g water, served at 19 °C; other pairs: Nariño 14 h 2.17 (a) and 2.00 (ab), Huila 22 h 1.67 (ab) and 1.17 (ab), Nariño 22 h 2.17 (a) and 1.50 (ab) | taste bitterness, 0 (not perceptible) to 5 (very strong) (3 experienced tasters, 3 replicates per sample; table gives means ± 95% confidence intervals, n=6); LSD letters in brackets | 1.50 (ab) | 1.00 (b) | Brewing compared more bitter Tested, not significant |
| Córdoba 2019 | 22 h vs 14 h (Huila, coarse-ground cold brew) Colombian Caturra (medium roast); cold brew by immersion at 20 °C, 60 g grounds to 700 g water, served at 19 °C; other pairs: Nariño coarse 2.17 (a) and 2.17 (a), Huila medium 1.17 (ab) and 1.00 (b), Nariño medium 1.50 (ab) and 2.00 (ab) | taste bitterness, 0 (not perceptible) to 5 (very strong) (3 experienced tasters, 3 replicates per sample; table gives means ± 95% confidence intervals, n=6); LSD letters in brackets | 1.67 (ab) | 1.50 (ab) | Brewing compared more bitter Tested, not significant |
| Córdoba 2019 | Reference (serving temperature and session differ): hot brew (French press) vs 4 cold brews (Huila) Colombian Caturra (medium roast); hot: French press with boiling water, 42.5 g grounds to 500 g water, served at 70 °C; cold: 20 °C, 60 g grounds to 700 g water, 4 versions of coarse or medium grind × 14 or 22 h, served at 19 °C; hot and cold brews rated in separate sessions | taste bitterness, 0 (not perceptible) to 5 (very strong) (3 experienced tasters, 3 replicates per sample; table gives means ± 95% confidence intervals, n=6); LSD letters in brackets | 2.00 (ab) | 1.00–1.67 (b to ab) | Brewing compared more bitter Tested, not significant |
| Córdoba 2019 | Reference (serving temperature and session differ): hot brew (French press) vs 4 cold brews (Nariño) Colombian Caturra (medium roast); hot: French press with boiling water, 42.5 g grounds to 500 g water, served at 70 °C; cold: 20 °C, 60 g grounds to 700 g water, 4 versions of coarse or medium grind × 14 or 22 h, served at 19 °C; hot and cold brews rated in separate sessions | taste bitterness, 0 (not perceptible) to 5 (very strong) (3 experienced tasters, 3 replicates per sample; table gives means ± 95% confidence intervals, n=6); LSD letters in brackets | 2.00 (ab) | 1.50–2.17 (ab to a) | Depends on the combination Tested, not significant |
| Chen 2023 | Ultra-high pressure: 500 MPa, 20 min vs conventional cold brew (atmospheric, 5 °C, 12 h) The same Ethiopian arabica, roasted medium (L value 24–26) and ground to 400–600 μm; 15 g grounds to 210 g water at 5 °C; for ultra-high pressure the bottle is placed in the unit; conventional cold brew at atmospheric pressure, 12 h in a refrigerator at 5 °C; all filtered, kept at 5 °C and served at 20 ± 3 °C; strength (TDS) 1.22% at 500 MPa and 1.21% for the conventional cold brew | 0 (none) to 15 (extremely intense) in 0.5 steps (6 Q Graders, 3 replicates); mean ± SD, significance letters in brackets | 5.20 ± 0.48 (b) | 6.75 ± 0.53 (a) | Comparison more bitter Tested, significant |
| Chen 2023 | Ultra-high pressure: 100 MPa vs 500 MPa (both 20 min) The same Ethiopian arabica, roasted medium (L value 24–26) and ground to 400–600 μm; 15 g grounds to 210 g water at 5 °C; for ultra-high pressure the bottle is placed in the unit; conventional cold brew at atmospheric pressure, 12 h in a refrigerator at 5 °C; all filtered, kept at 5 °C and served at 20 ± 3 °C; 200 MPa 5.88 (ab), 300 MPa 5.38 (b), 400 MPa 5.25 (b) | 0 (none) to 15 (extremely intense) in 0.5 steps (6 Q Graders, 3 replicates); mean ± SD, significance letters in brackets | 6.00 ± 0.66 (ab) | 5.20 ± 0.48 (b) | Brewing compared more bitter Tested, not significant |
| Chen 2023 | Ultra-high pressure: 300 MPa, 30 min vs 10 min The same Ethiopian arabica, roasted medium (L value 24–26) and ground to 400–600 μm; 15 g grounds to 210 g water at 5 °C; for ultra-high pressure the bottle is placed in the unit; conventional cold brew at atmospheric pressure, 12 h in a refrigerator at 5 °C; all filtered, kept at 5 °C and served at 20 ± 3 °C; 15 min 4.85 (bc), 20 min 5.38 (bc), 25 min 5.50 (abc); strength (TDS) 1.11% at 10 min and 1.24% at 30 min | 0 (none) to 15 (extremely intense) in 0.5 steps (6 Q Graders, 3 replicates); mean ± SD, significance letters in brackets | 5.88 ± 0.72 (ab) | 4.25 ± 0.68 (c) | Brewing compared more bitter Tested, significant |
| Shi 2024 | Ultra-high pressure: conventional cold brew vs 300 MPa, 20 min (light roast) Ethiopian arabica, light roast (Agtron colour zone 73); 15 g grounds to 210 mL water; ultra-high pressure at 300 MPa for 20 min (water temperature not given) and conventional cold brew at 4 °C for 12 h; both filtered, kept at 4 °C and served at 20 ± 3 °C; strength (TDS) 1.11% for the conventional cold brew and 1.10% under ultra-high pressure (no significant difference within the same roast) | scale range not given (12 Q Arabica Graders); mean ± SD, significance letters in brackets | 3.00 ± 0.28 (d) | 2.00 ± 0.10 (e) | Brewing compared more bitter Tested, significant |
| Shi 2024 | Ultra-high pressure: conventional cold brew vs 300 MPa, 20 min (medium roast) Ethiopian arabica, medium roast (Agtron colour zone 62); 15 g grounds to 210 mL water; ultra-high pressure at 300 MPa for 20 min (water temperature not given) and conventional cold brew at 4 °C for 12 h; both filtered, kept at 4 °C and served at 20 ± 3 °C; strength (TDS) 1.21% for the conventional cold brew and 1.18% under ultra-high pressure (no significant difference within the same roast) | scale range not given (12 Q Arabica Graders); mean ± SD, significance letters in brackets | 4.30 ± 0.42 (c) | 3.35 ± 0.30 (d) | Brewing compared more bitter Tested, significant |
| Shi 2024 | Ultra-high pressure: conventional cold brew vs 300 MPa, 20 min (dark roast) Ethiopian arabica, dark roast (Agtron colour zone 53); 15 g grounds to 210 mL water; ultra-high pressure at 300 MPa for 20 min (water temperature not given) and conventional cold brew at 4 °C for 12 h; both filtered, kept at 4 °C and served at 20 ± 3 °C; strength (TDS) 1.30% for the conventional cold brew and 1.28% under ultra-high pressure (no significant difference within the same roast) | scale range not given (12 Q Arabica Graders); mean ± SD, significance letters in brackets | 6.50 ± 0.42 (a) | 5.50 ± 0.40 (b) | Brewing compared more bitter Tested, significant |
Conditions and results, study by study
For each study, the conditions for the beans, roast, brewing and serving, and who rated them, and the results. Each row is one study × one measured result, so different comparisons in the same study are separate rows. Under each row we give where it is in the paper and whether it can be used to judge whether changing only the brewing of the same beans changes the intensity of bitterness people perceive, with the reason. Nothing is left blank: "not given", "not read" and "not applicable" are kept apart, and values from figures we did not obtain are marked as such.
Cai 2022 sensory only
Comparative Profiling of Hot and Cold Brew Coffee Flavor Using Chromatographic and Sensory Approaches. Foods 11(19):2968 — doi:10.3390/foods11192968
- How deeply checked
- Checked in full text — Full text via Europe PMC (sections 2.1, 2.2, 2.6 and 2.7, the sensory paragraphs of the results, Table 4) Checked by: LLM (one extracted, another checked against the source; no human check), 2026-10-22.
- Samples
- A commercial medium-dark arabica blend (Costa Rica, Colombia, Indonesia); 2.0 kg from the same batch mixed before use
- Processing and cooking
- Commercially roasted beans (described only as medium-dark; roasting conditions not given)
- How it was served or eaten
- Ground to 30–60 mesh. Hot: coffee machine, 60 g grounds to 600 g water, outlet 80 ± 3 °C, about 6 min, paper filter. Cold: 300-mesh bag, same 60 g and 600 g, 20 ± 3 °C for 8 h, then rinsed with another 120 g water and combined. Both diluted to 2% TDS and rated at 20 ± 3 °C after refrigeration
- Who rated, on what scale
- 7 university students (aged 19–22, 2 men and 5 women, drinking coffee for 3 years or more, 3 or more training rounds); 1–5 points per the methods; Student t-test (95%); a separate triangle test with 48 people
| Measured | Comparison and condition | Brewing compared | Comparison | Unit | Direction / test |
|---|---|---|---|---|---|
| Bitterness intensity (sensory) | Hot brew (coffee machine) vs cold brew (8 h at room temperature) Same batch of a medium-dark arabica blend. Hot: 60 g grounds to 600 g water, outlet 80 ± 3 °C, about 6 min. Cold: same 60 g and 600 g, 20 ± 3 °C for 8 h, then rinsed with another 120 g water and combined. Both diluted to 2% strength (TDS) and served at 20 ± 3 °C | 4.14 ± 0.69 | 2.43 ± 0.53 | 1–5 points per the methods (mean ± SD of 7 trained panellists); the sweetness value of 0.71 in the same table falls outside 1–5, so the lower end of the scale is inconsistent within the source | Brewing compared more bitter Tested, significant |
| Student t-test, p<0.05 (mark in Table 4); astringency was also higher hot (2.86 and 2.00, significant); sourness was 2.86 and 3.29, with no significant difference detected; in a separate triangle test with 48 people, 25 picked the odd sample, above the critical value of 23 Where: Sections 2.2, 2.6 and 2.7; Table 4 For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Usable — Same batch of beans with only the brewing changed, strength and serving temperature matched, rated by 7 trained panellists; hot was a coffee machine and cold a room-temperature immersion, so temperature, time and device differ together (a comparison of whole brewing methods) | |||||
Heo 2019 sensory only
Cold Brew Coffee: Consumer Acceptability and Characterization Using the Check-All-That-Apply (CATA) Method. Foods 8(8):344 — doi:10.3390/foods8080344
- How deeply checked
- Checked in full text — Full text via Europe PMC (sections 2.1, 2.1.1, 2.2 and 2.4, Tables 1 and 3); the CATA results and the commercial products are not used in this table Checked by: LLM (one extracted, another checked against the source; no human check), 2026-10-22.
- Samples
- The same beans (or the same product), 4 kinds: a blend (city roast; Brazil, Colombia, Ethiopia, Guatemala), Colombia (Nariño, medium), Folgers classic roast (medium) and Kenya AA (medium); besides these 8 samples, Table 3 has 6 rows (commercial and coffee-shop cold brews, including a blind duplicate), not tabled here; section 2.1 gives 13 samples including the duplicate, which does not match the number of table rows
- Processing and cooking
- Commercially roasted (roasting conditions not given)
- How it was served or eaten
- Medium grind, 7 g grounds to 100 mL water for both. Cold brew: 9 h refrigerated at 4 °C, filtered through metal and paper, used within 22 h of brewing. Coffee maker: paper filter, cooled in an ice bath for 30 min and refrigerated, used within 10 h of brewing. All served at about 8 ± 2 °C, 50 mL
- Who rated, on what scale
- 120 untrained consumers; bitterness intensity on a 9-point category scale and bitterness liking on a separate 9-point hedonic scale; 2 sessions; one-way ANOVA and Fisher LSD
| Measured | Comparison and condition | Brewing compared | Comparison | Unit | Direction / test |
|---|---|---|---|---|---|
| Bitterness intensity (sensory) | Coffee maker vs cold brew (blend) The same beans brewed as cold brew (9 h at 4 °C, kept refrigerated, used within 22 h) and in a coffee maker (cooled 30 min in an ice bath and refrigerated, used within 10 h); 7 g grounds to 100 mL water and medium grind for both; served at about 8 ± 2 °C, 50 mL. Beans: city-roast blend (Brazil, Colombia, Ethiopia, Guatemala) | 4.9 (d) | 4.9 (d) | intensity on a 9-point category scale (mean of 120 untrained consumers); LSD letters in brackets (values sharing a letter do not differ significantly) | About the same Tested, not significant |
| One-way ANOVA over 14 samples including commercial cold brews (bitterness intensity p<0.0001) and Fisher LSD (LSD 0.47); separate from the bitterness-liking column; serving temperature was the same, but brewing temperature and the cooling and holding before serving differ; equal means with the same letter, so no significant difference was detected; this does not show that there is no difference Where: Sections 2.1, 2.1.1, 2.2 and 2.4; Tables 1 and 3 (Bitterness Intensity column) For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Limited — Same beans brewed 2 ways and served at the same temperature, but rated by untrained consumers, with no record of matching strength, and cooling and holding times differ | |||||
| Bitterness intensity (sensory) | Coffee maker vs cold brew (Colombia) The same beans brewed as cold brew (9 h at 4 °C, kept refrigerated, used within 22 h) and in a coffee maker (cooled 30 min in an ice bath and refrigerated, used within 10 h); 7 g grounds to 100 mL water and medium grind for both; served at about 8 ± 2 °C, 50 mL. Beans: Colombia (Nariño), medium roast | 4.9 (d) | 4.6 (de) | intensity on a 9-point category scale (mean of 120 untrained consumers); LSD letters in brackets (values sharing a letter do not differ significantly) | Brewing compared more bitter Tested, not significant |
| One-way ANOVA over 14 samples including commercial cold brews (bitterness intensity p<0.0001) and Fisher LSD (LSD 0.47); separate from the bitterness-liking column; serving temperature was the same, but brewing temperature and the cooling and holding before serving differ; the direction shows which published mean is larger; the two share the letter d, so no significant difference was detected Where: Sections 2.1, 2.1.1, 2.2 and 2.4; Tables 1 and 3 (Bitterness Intensity column) For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Limited — Same beans brewed 2 ways and served at the same temperature, but rated by untrained consumers, with no record of matching strength, and cooling and holding times differ | |||||
| Bitterness intensity (sensory) | Coffee maker vs cold brew (Folgers) The same beans brewed as cold brew (9 h at 4 °C, kept refrigerated, used within 22 h) and in a coffee maker (cooled 30 min in an ice bath and refrigerated, used within 10 h); 7 g grounds to 100 mL water and medium grind for both; served at about 8 ± 2 °C, 50 mL. Commercial Folgers classic roast (medium) | 6.8 (a) | 5.8 (bc) | intensity on a 9-point category scale (mean of 120 untrained consumers); LSD letters in brackets (values sharing a letter do not differ significantly) | Brewing compared more bitter Tested, significant |
| One-way ANOVA over 14 samples including commercial cold brews (bitterness intensity p<0.0001) and Fisher LSD (LSD 0.47); separate from the bitterness-liking column; serving temperature was the same, but brewing temperature and the cooling and holding before serving differ; the letters do not overlap, so the coffee maker brew was significantly higher Where: Sections 2.1, 2.1.1, 2.2 and 2.4; Tables 1 and 3 (Bitterness Intensity column) For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Limited — Same beans brewed 2 ways and served at the same temperature, but rated by untrained consumers, with no record of matching strength, and cooling and holding times differ | |||||
| Bitterness intensity (sensory) | Coffee maker vs cold brew (Kenya) The same beans brewed as cold brew (9 h at 4 °C, kept refrigerated, used within 22 h) and in a coffee maker (cooled 30 min in an ice bath and refrigerated, used within 10 h); 7 g grounds to 100 mL water and medium grind for both; served at about 8 ± 2 °C, 50 mL. Beans: Kenya AA, medium roast | 5.0 (d) | 4.4 (ef) | intensity on a 9-point category scale (mean of 120 untrained consumers); LSD letters in brackets (values sharing a letter do not differ significantly) | Brewing compared more bitter Tested, significant |
| One-way ANOVA over 14 samples including commercial cold brews (bitterness intensity p<0.0001) and Fisher LSD (LSD 0.47); separate from the bitterness-liking column; serving temperature was the same, but brewing temperature and the cooling and holding before serving differ; the letters do not overlap, so the coffee maker brew was significantly higher Where: Sections 2.1, 2.1.1, 2.2 and 2.4; Tables 1 and 3 (Bitterness Intensity column) For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Limited — Same beans brewed 2 ways and served at the same temperature, but rated by untrained consumers, with no record of matching strength, and cooling and holding times differ | |||||
Lopane 2024 sensory only
An investigation of the shelf life of cold brew coffee and the influence of extraction temperature using chemical, microbial, and sensory analysis. Food Science & Nutrition 12(2):985-996 — doi:10.1002/fsn3.3812
- How deeply checked
- Checked in full text — Full text via Europe PMC (methods, results and discussion, conclusions); Figs 4 and 5 with the bitterness values not obtained Checked by: LLM (one extracted, another checked against the source; no human check), 2026-10-22.
- Samples
- One lot of natural-process Brazilian arabica (Cerrado Mineiro), variety Acaiá Cerrado
- Processing and cooking
- Roast degree Gourmet 72.9 (brewed 6 days after roasting); same grind
- How it was served or eaten
- Water standardised by adding minerals. Cold: immersion at 4 °C for 4 h, coffee to water 1:14. Ambient: the same at 25 °C. Hot: drip at about 97 °C, 1:18, flash-chilled to 7 °C in an ice bath. Bottled and stored at 7 °C, served at 7 °C on days 0, 7, 14, 28 and 42
- Who rated, on what scale
- 5 coffee professionals (cupping at least once a week); WCR lexicon, intensity 0–15, 3 replicates; mixed model and Fisher LSD (5%)
| Measured | Comparison and condition | Brewing compared | Comparison | Unit | Direction / test |
|---|---|---|---|---|---|
| Bitterness intensity (sensory) | Hot drip vs 4 °C immersion (storage day 0) The same Brazilian arabica (Acaiá, roast degree Gourmet 72.9), same grind, mineral-standardised water. Hot: drip at about 97 °C, coffee to water 1:18, flash-chilled to 7 °C in an ice bath. Cold: immersion at 4 °C for 4 h, 1:14; there is also a 25 °C immersion under the same conditions. Bottled, stored at 7 °C and served at 7 °C | Figure not obtained, so values unconfirmed | Figure not obtained, so values unconfirmed | intensity from 0 (none) to 15 (5 coffee professionals, 3 replicates) | Brewing compared more bitter Tested, significant |
| From the text: the 4 °C and 25 °C immersions were significantly less bitter than the hot drip (mixed model and Fisher LSD, 5%); values are only in Fig. 5. Figure not obtained, so values unconfirmed. The hot brew also differs in coffee-to-water ratio (1:18 and 1:14) and flash chilling, so this is not a comparison of temperature alone (the authors note the same); strength (TDS) was highest for the 25 °C immersion and lowest for the hot drip Where: Methods; Results and Discussion, "Effect of extraction temperature and method—Sensory profile"; Conclusions; Fig. 5 (not obtained) For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Limited — Same beans with the brewing changed, served at the same temperature and rated by professionals, but values are only in a figure and unconfirmed, and the hot brew also differs in coffee-to-water ratio and flash chilling; 5 people rated | |||||
| Bitterness intensity (sensory) | Hot drip vs 4 °C immersion (storage day 42) The same Brazilian arabica (Acaiá, roast degree Gourmet 72.9), same grind, mineral-standardised water. Hot: drip at about 97 °C, coffee to water 1:18, flash-chilled to 7 °C in an ice bath. Cold: immersion at 4 °C for 4 h, 1:14; there is also a 25 °C immersion under the same conditions. Bottled, stored at 7 °C and served at 7 °C | Figure not obtained, so values unconfirmed | Figure not obtained, so values unconfirmed | intensity from 0 (none) to 15 (5 coffee professionals, 3 replicates) | Brewing compared more bitter Tested, significant |
| From the text: the 4 °C and 25 °C immersions were significantly less bitter than the hot drip (mixed model and Fisher LSD, 5%); values are only in Fig. 5. Figure not obtained, so values unconfirmed. The hot brew also differs in coffee-to-water ratio (1:18 and 1:14) and flash chilling, so this is not a comparison of temperature alone (the authors note the same); strength (TDS) was highest for the 25 °C immersion and lowest for the hot drip; the change in bitterness with storage is described inconsistently within the same text (one sentence says day 0 was lower than day 42, another that bitterness decreased with storage), and Fig. 4 was not obtained, so it is not used on this page Where: Results and Discussion, "Effect of extraction temperature and method—Sensory profile"; Conclusions; Fig. 5 (not obtained) For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Limited — Same beans with the brewing changed, served at the same temperature and rated by professionals, but values are only in a figure and unconfirmed, and the hot brew also differs in coffee-to-water ratio and flash chilling; 5 people rated | |||||
Batali 2022 sensory only
Sensory Analysis of Full Immersion Coffee: Cold Brew Is More Floral, and Less Bitter, Sour, and Rubbery Than Hot Brew. Foods 11(16):2440 — doi:10.3390/foods11162440
- How deeply checked
- Checked in full text — Full text via Europe PMC (Tables 1–3 and text); supplementary Fig. S1 not viewed Checked by: LLM (one extracted, another checked against the source; no human check), 2026-10-21.
- Samples
- 3 arabicas (El Salvador, Ethiopia, Sumatra)
- Processing and cooking
- Agtron Gourmet values light 58, medium 48, dark 38 (medium and dark are averages); temperature, time and weight loss not given
- How it was served or eaten
- Full immersion, 100 g grounds to 500 g water: 4 °C (22 °C water added, then refrigerated 24–36 h), 22 °C (12–14 h) or 92 °C (left to cool at room temperature, 1–2 h), until extraction stopped; all diluted to 2% strength and served at 4 °C
- Who rated, on what scale
- Rated at home: 20 trained, 17 evaluated, 14 completed all sessions, and 4 with poor reproducibility excluded, leaving 10 analysed; 0–100 line scale, 3 replicates; ANOVA
| Measured | Comparison and condition | Brewing compared | Comparison | Unit | Direction / test |
|---|---|---|---|---|---|
| Bitterness intensity (sensory) | Overall result (pooled over origins and roasts): 92 °C vs 4 °C immersion (9 coffees) 9 coffees (3 arabica origins, each at 3 roast levels) immersed at 100 g grounds to 500 g water at 4 °C (22 °C water poured, then refrigerated 24–36 h), 22 °C (12–14 h) or 92 °C water (left to cool at room temperature, 1–2 h) until extraction stopped; all diluted to 2% strength (TDS) and served at 4 °C; values are the main effect of temperature pooled over origins and roasts | Figure not obtained, so values unconfirmed | Figure not obtained, so values unconfirmed | 0–100 line scale (rated at home, 10 panellists analysed, 3 replicates) | Brewing compared more bitter Tested, significant |
| From the text: in the 5-factor ANOVA the main effect of temperature on bitterness was F=6.06, and by LSD 92 °C was significantly higher than 4 °C and 22 °C; this is pooled over origins and roasts and does not show significance for each of the 9 coffees; the means are only in Fig. 4. Figure not obtained, so values unconfirmed Where: Sections 2.3, 2.5 and 2.6; Table 3; section 3.3; Fig. 4 (not obtained) For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Limited — Immersion temperatures compared with strength and serving temperature matched, but values are only in a figure and unconfirmed, and the result is pooled over 9 coffees | |||||
| Bitterness intensity (sensory) | Overall result (pooled over origins and roasts): 22 °C vs 4 °C immersion (9 coffees) 9 coffees (3 arabica origins, each at 3 roast levels) immersed at 100 g grounds to 500 g water at 4 °C (22 °C water poured, then refrigerated 24–36 h), 22 °C (12–14 h) or 92 °C water (left to cool at room temperature, 1–2 h) until extraction stopped; all diluted to 2% strength (TDS) and served at 4 °C; values are the main effect of temperature pooled over origins and roasts | Figure not obtained, so values unconfirmed | Figure not obtained, so values unconfirmed | 0–100 line scale (rated at home, 10 panellists analysed, 3 replicates) | Direction not given Tested, not significant |
| From the text: in the overall temperature comparison no significant difference was detected between 4 °C and 22 °C (LSD); this does not show they were the same for each of the 9 coffees; the means are only in Fig. 4. Figure not obtained, so values unconfirmed Where: Section 3.3; Fig. 4 (not obtained) For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Limited — As above; a result where the overall comparison detected no significant difference | |||||
Liang 2024 sensory only
Sensory analysis of the flavor profile of full immersion hot, room temperature, and cold brewed coffee over time. Scientific Reports 14 — doi:10.1038/s41598-024-69867-6
- How deeply checked
- Checked in full text — Full text via Europe PMC and images of Table 1 and Fig. 5; Fig. 5 values estimated by eye Checked by: LLM (one extracted, another checked against the source; no human check), 2026-10-21.
- Samples
- One Central American blend
- Processing and cooking
- Agtron Gourmet values light 71.8 and dark 41.8 (the authors call the darker one the dark end of medium-dark); temperature and time not given
- How it was served or eaten
- Full immersion, 100 g grounds to 500 g water; 5 time points, 0–24 h at 4 and 22 °C and 0–60 min at 92 °C; concentrate diluted 2-fold (strength not equalised); served at about 6 °C, warming to about 12 °C while tasted
- Who rated, on what scale
- 14 trained panellists; 0–100 line scale, 3 replicates; ANOVA and LSD
| Measured | Comparison and condition | Brewing compared | Comparison | Unit | Direction / test |
|---|---|---|---|---|---|
| Bitterness intensity (sensory) | 92 °C and 22 °C vs 4 °C immersion (darker roast) The darker roast of the same Central American blend (Agtron 41.8); 100 g grounds to 500 g water, 5 time points over 0–24 h for 4 °C water (refrigerated) and 22 °C water (room temperature) and over 0–60 min for 92 °C water (left at room temperature); concentrate diluted 2-fold (strength not equalised); served at about 6 °C, warming to about 12 °C while tasted; values compare the pooled 5 time points | Figure not obtained, so values unconfirmed | Figure not obtained, so values unconfirmed | 0–100 line scale (14 trained panellists, 3 replicates) | Brewing compared more bitter Tested, significant |
| For the darker roast, bitterness increased significantly as temperature rose from 4 °C to 22 °C to 92 °C; in the cold immersions bitterness changed significantly over time (Fig. 7). From the text (ANOVA and LSD); the comparison pools the 5 time points and does not show the same difference at each time; the text gives mean strength (TDS) before dilution at the final time point as 3.97% at 92 °C, 3.93% at 22 °C and 3.52% at 4 °C (a description of the whole study, not values for this roast alone), and the text says bitterness correlated positively with strength; time 0 is a "rinsed" sample filtered at once; values are only in Fig. 5, which is not in this question's store, so values are unconfirmed (a by-eye reading of Fig. 5 is recorded in the Codex check of 2026-10-21 for the dark-roast question, but for roast comparisons; the bitterness values by temperature for this question were not checked) Where: Methods; Results, "Effects of roast, temperature, and time"; Figs 5 and 7 (not obtained) For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Limited — Same beans and roast with immersion temperature varied, but strength was not equalised, values are only in a figure and unconfirmed, and the comparison pools 5 time points | |||||
| Bitterness intensity (sensory) | 92 °C vs 4 °C and 22 °C immersion (lighter roast) The lighter roast of the same Central American blend (Agtron 71.8); 100 g grounds to 500 g water, 5 time points over 0–24 h for 4 °C water (refrigerated) and 22 °C water (room temperature) and over 0–60 min for 92 °C water (left at room temperature); concentrate diluted 2-fold (strength not equalised); served at about 6 °C, warming to about 12 °C while tasted; values compare the pooled 5 time points | Figure not obtained, so values unconfirmed | Figure not obtained, so values unconfirmed | 0–100 line scale (14 trained panellists, 3 replicates) | Brewing compared more bitter Tested, significant |
| For the lighter roast, bitterness at both 4 °C and 22 °C was significantly lower than at 92 °C. From the text (ANOVA and LSD); the comparison pools the 5 time points and does not show the same difference at each time; the text gives mean strength (TDS) before dilution at the final time point as 3.97% at 92 °C, 3.93% at 22 °C and 3.52% at 4 °C (a description of the whole study, not values for this roast alone), and the text says bitterness correlated positively with strength; time 0 is a "rinsed" sample filtered at once; values are only in Fig. 5, which is not in this question's store, so values are unconfirmed (a by-eye reading of Fig. 5 is recorded in the Codex check of 2026-10-21 for the dark-roast question, but for roast comparisons; the bitterness values by temperature for this question were not checked) Where: Results, "Effects of roast, temperature, and time"; Fig. 5 (not obtained) For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Limited — Same beans and roast with immersion temperature varied, but strength was not equalised, values are only in a figure and unconfirmed, and the comparison pools 5 time points | |||||
| Bitterness intensity (sensory) | 22 °C vs 4 °C immersion (lighter roast) The lighter roast of the same Central American blend (Agtron 71.8); 100 g grounds to 500 g water, 5 time points over 0–24 h for 4 °C water (refrigerated) and 22 °C water (room temperature) and over 0–60 min for 92 °C water (left at room temperature); concentrate diluted 2-fold (strength not equalised); served at about 6 °C, warming to about 12 °C while tasted; values compare the pooled 5 time points | Figure not obtained, so values unconfirmed | Figure not obtained, so values unconfirmed | 0–100 line scale (14 trained panellists, 3 replicates) | Direction not given Tested, not significant |
| For the lighter roast, no significant difference was detected between 4 °C and 22 °C; this does not show they were the same. From the text (ANOVA and LSD); the comparison pools the 5 time points and does not show the same difference at each time; the text gives mean strength (TDS) before dilution at the final time point as 3.97% at 92 °C, 3.93% at 22 °C and 3.52% at 4 °C (a description of the whole study, not values for this roast alone), and the text says bitterness correlated positively with strength; time 0 is a "rinsed" sample filtered at once; values are only in Fig. 5, which is not in this question's store, so values are unconfirmed (a by-eye reading of Fig. 5 is recorded in the Codex check of 2026-10-21 for the dark-roast question, but for roast comparisons; the bitterness values by temperature for this question were not checked) Where: Results, "Effects of roast, temperature, and time"; Fig. 5 (not obtained) For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Limited — Same beans and roast with immersion temperature varied, but strength was not equalised, values are only in a figure and unconfirmed, and the comparison pools 5 time points | |||||
Batali 2020 sensory only
Brew temperature, at fixed brew strength and extraction, has little impact on the sensory profile of drip brew coffee. Scientific Reports — doi:10.1038/s41598-020-73341-4
- How deeply checked
- Checked in full text — Full text via Europe PMC (results, Table 2, discussion, methods); Fig. 3 with the bitterness means and the supplementary tables not obtained Checked by: LLM (one extracted, another checked against the source; no human check), 2026-10-22.
- Samples
- Washed arabica from Marcala, Honduras (a mix of several varieties including Bourbon, Catuai and Caturra)
- Processing and cooking
- Roasted on the same day in several 29 kg batches (11 min 29 s total, first crack at 9 min), then mixed after roasting
- How it was served or eaten
- Commercial drip brewer (flat-bottomed basket, paper filter); at brew temperatures of 87, 90 and 93 °C, 9 targets of strength (TDS about 1, 1.25 and 1.5%) and extraction were reached by adjusting grind, dose and flow (27 types); water made to the industry standard by dissolving salts; served after cooling to 64–66 °C
- Who rated, on what scale
- 12 trained panellists (aged 19–30, trained 3 times a week for 4 weeks); 15 cm line scale converted to 0–100, 3 replicates; MANOVA, 5-factor ANOVA with LSD, response surface
| Measured | Comparison and condition | Brewing compared | Comparison | Unit | Direction / test |
|---|---|---|---|---|---|
| Bitterness intensity (sensory) | Drip at 93 °C vs 87 °C (and 90 °C), strength and extraction matched Honduran arabica roasted on the same day and mixed; commercial drip brewer (flat-bottomed, paper filter); at 87, 90 and 93 °C, 9 targets of strength (TDS) and extraction were reached by adjusting grind, dose and flow, matched across temperatures; served after cooling to 64–66 °C | Figure not obtained, so values unconfirmed | Figure not obtained, so values unconfirmed | 15 cm line converted to 0–100 (12 trained panellists, 3 replicates) | Direction not given Tested, not significant |
| In the 5-factor ANOVA, bitterness was among the attributes that varied significantly with strength (TDS) alone, and the only attribute that varied significantly with brew temperature was nutty flavour; F=1.25 and p=0.11 for temperature in Table 2 are from the MANOVA across all sensory attributes, not a test of bitterness alone; no significant difference was detected across these 3 temperatures, which does not show that bitterness was the same; means by temperature are unconfirmed; the text says Fig. 3 leaves entries with no significant difference blank (the figure was not obtained) Where: Results, "Multivariate and univariate analyses of variance"; Table 2; Fig. 3 (not obtained); Discussion For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Usable — Same beans; the effect of brew temperature was compared after adjusting grind, dose and flow to match strength and extraction, with serving temperature matched and 12 trained panellists rating; means by temperature unconfirmed | |||||
| Bitterness intensity (sensory) | Strength (TDS) about 1.5% vs about 1% (pooled over temperature and extraction) Honduran arabica roasted on the same day and mixed; commercial drip brewer (flat-bottomed, paper filter); at 87, 90 and 93 °C, 9 targets of strength (TDS) and extraction were reached by adjusting grind, dose and flow, matched across temperatures; served after cooling to 64–66 °C; the middle level was about 1.25% | Figure not obtained, so values unconfirmed | Figure not obtained, so values unconfirmed | 15 cm line converted to 0–100 (12 trained panellists, 3 replicates) | Brewing compared more bitter Tested, significant |
| From the text: bitterness increased linearly with strength (TDS) and had no significant relation with extraction (ANOVA and response surface); strength was made by changing grind, dose and flow, so this is not the effect of the coffee-to-water ratio alone; the means are in Figs 3 and 6. Figure not obtained, so values unconfirmed Where: Results, response surface paragraphs; Fig. 6 (not obtained); Discussion For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Limited — Same beans with strength varied, but strength was made through grind, dose and flow, so which part of the brewing mattered cannot be separated | |||||
Izumi and Takaya 2008 sensory only
コーヒーの味に及ぼす抽出条件およびクロロゲン酸量の影響. 日本調理科学会誌 41(4):257-261 — doi:10.11402/cookeryscience1995.41.4_257
- How deeply checked
- Checked in full text — Full 5-page PDF from J-STAGE (samples and methods, results and discussion part 1, Figs 1 and 2); figure values are approximate readings by eye from page images Checked by: LLM (one extracted, another checked against the source; no human check), 2026-10-22.
- Samples
- 3 commercially roasted coffees: A (Mocha Sidamo), B (Colombia), C (Mocha, from another shop); bought at specialty shops in Sendai
- Processing and cooking
- Commercially roasted (roast degree not given); ground just before brewing (setting 4 (560–450 μm) for the siphon, setting 9 (710–560 μm) otherwise)
- How it was served or eaten
- Temperature: paper drip, 30 g grounds to 450 mL water, 20 s bloom, 4 min total, distilled water, at 85, 90 and 95 °C. Equipment: paper drip (90 °C), electric coffee maker (same amounts) and siphon (20 g grounds to 300 mL water). Water: distilled, a commercial soft water (hardness 30) and a commercial hard water (hardness 1551), paper drip at 90 °C. One person brewed everything; kept in vacuum flasks so samples were at as similar a temperature as possible when rated
- Who rated, on what scale
- 12–14 women (aged 21–55; students, assistants and teachers in food courses, experienced in sensory evaluation, who like and regularly drink coffee); 5-point scoring from 1 to 5 (5 = very strong); two-way ANOVA
| Measured | Comparison and condition | Brewing compared | Comparison | Unit | Direction / test |
|---|---|---|---|---|---|
| Bitterness intensity (sensory) | Paper drip at 95 °C vs 85 °C (and 90 °C) (3 coffees) 3 commercially roasted coffees (A Mocha Sidamo, B Colombia, C Mocha); paper drip, 30 g grounds to 450 mL water, 20 s bloom, 4 min total, distilled water; kept in vacuum flasks so samples were at as similar a temperature as possible when rated | No values given | No values given | 5-point scoring from 1 to 5 (5 = very strong); 12–14 women | Direction not given Tested, not significant |
| The text says that for all 3 coffees there was no difference in any attribute except colour (two-way ANOVA); no significant difference was detected across these 3 temperatures, which does not show that bitterness was the same Where: Sample preparation, brewing methods (i) water temperature; Results and Discussion part 1 (p.258) For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Limited — Same beans with only water temperature changed, but no values are shown, only the text | |||||
| Bitterness intensity (sensory) | Siphon vs paper drip (coffee A; grind also differs) Commercially roasted coffee A (Mocha Sidamo, roast degree not given); paper drip (90 °C, 30 g grounds to 450 mL water), electric coffee maker (same amounts) and siphon (20 g grounds to 300 mL water); grind setting 4 (560–450 μm) for the siphon and 9 (710–560 μm) otherwise; the coffee-to-water ratio is 1:15 for both paper drip and siphon; kept in vacuum flasks so samples were at as similar a temperature as possible when rated; the coffee maker was about 2.5 | about 3.3 | about 2.3 | 5-point scoring from 1 to 5 (5 = very strong); approximate values read by eye from the figure; 12–14 women (Fig. 1) | Brewing compared more bitter Tested; significance reported, but which pair differs is unclear |
| Bitterness in Fig. 1 is marked p<0.01 (comparison of the 3 devices); the text says the siphon was significantly more bitter, but no comparison showing which 2 devices differ is given; grind and the amounts of coffee and water differ as well as the device; for coffee B bitterness is marked p<0.01 and for coffee C p<0.05, and in the figure the coffee maker is lower for both Where: Sample preparation, brewing methods (ii) equipment; Results and Discussion part 1 (p.258); Fig. 1 (p.259) For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Limited — Same beans with the device changed, but grind and amounts also differ, which pair differs is not shown, and values are approximate readings from a figure | |||||
| Bitterness intensity (sensory) | Commercial hard water vs distilled water (coffee B; a comparison of the commercial waters used) Commercially roasted coffee B (Colombia, roast degree not given); paper drip, 90 °C, 30 g grounds to 450 mL water; waters were distilled, a commercial soft water (hardness 30) and a commercial hard water (Contrex, hardness 1551, unit not given); the soft water was about 3.2 | about 3.7 | about 2.8 | 5-point scoring from 1 to 5 (5 = very strong); approximate values read by eye from the figure; 12–14 women (Fig. 2) | Brewing compared more bitter Tested; significance reported, but which pair differs is unclear |
| Bitterness in Fig. 2 is marked p<0.01 (comparison of the 3 waters); the text says bitterness and body were significantly stronger with hard water, but which 2 waters differ is not shown; this compares the commercial waters used, whose composition (calcium, magnesium, bicarbonate and so on) is not given, so it cannot be called an effect of hardness alone; coffee A has no mark for bitterness (the marks are for sourness and colour), and coffee C has none for bitterness either Where: Sample preparation, brewing methods (iii) water; Results and Discussion part 1 (p.258); Fig. 2 (p.259) For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Limited — Same beans with only the water changed, but these are commercial waters with no composition given, which pair differs is not shown, and values are approximate readings from a figure | |||||
Vezzulli 2022 sensory only
Metabolomics Combined with Sensory Analysis Reveals the Impact of Different Extraction Methods on Coffee Beverages from Coffea arabica and Coffea canephora var. Robusta. Foods 11(6):807 — doi:10.3390/foods11060807
- How deeply checked
- Checked in full text — Read in the full text on PMC and its Table 2 Checked by: LLM (one extracted, another checked against the source; no human check), 2026-10-20.
- Samples
- One commercial blend of 100% arabica (Brazil, natural) and one of 100% robusta (India, natural); different countries of origin; one sample each
- Processing and cooking
- Roasted by the same roaster with the same roasting process (no roast-degree value reported)
- How it was served or eaten
- 4 methods: espresso (coffee to beverage 1:2 by weight, about 16 g, 92 °C), filter (50 g/L), moka (76 g/L) and Neapolitan pot (72 g/L)
- Who rated, on what scale
- 6 trained panelists; scale from 0 (absent) to 9 (very intense); the table gives the panel median
| Measured | Comparison and condition | Brewing compared | Comparison | Unit | Direction / test |
|---|---|---|---|---|---|
| Bitterness intensity (sensory) | Espresso vs filter (arabica) 100% Brazilian arabica (same roaster and roasting process, roast degree not given); espresso (92 °C, coffee to beverage 1:2, about 16 g) and filter from a home drip machine (50 g/L); the same beans as moka (76 g/L) 3.4 and Neapolitan pot (72 g/L) 3.1; espresso and filter rated in the same session, moka and Neapolitan in another | 4.0 | 1.2 | median on 0 (absent) to 9 (very intense) (6 trained panellists) | Brewing compared more bitter Test method described; result for this comparison not reported |
| The direction shows which published median is larger. ANOVA was carried out, but no result of the bitterness comparison is reported (bitterness is not in the list of significant attributes); no p value or pairwise comparison for bitterness is given; coffee-to-water ratios also differ by method; the abstract says bitterness was perceived as stronger in espresso, which matches the medians, but no support from a significance test is shown Where: Sections 2.2 and 2.3; Table 2; section 3.3 For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Limited — Same beans with the method changed, but ratios also differ, there is no support for a significant difference in bitterness, and values are medians of 6 people | |||||
| Bitterness intensity (sensory) | Espresso vs filter (robusta) 100% Indian robusta (same roaster and roasting process), brewed as for the arabica; the same beans as moka 3.3 and Neapolitan pot 2.9; arabica versus robusta belongs to another page, and only brewing methods within the same robusta are compared here | 4.7 | 3.5 | median on 0 (absent) to 9 (very intense) (6 trained panellists) | Brewing compared more bitter Test method described; result for this comparison not reported |
| The direction shows which published median is larger. ANOVA was carried out, but no result of the bitterness comparison is reported (bitterness is not in the list of significant attributes); no p value or pairwise comparison for bitterness is given; coffee-to-water ratios also differ by method Where: Sections 2.2 and 2.3; Table 2; section 3.3 For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Limited — Same beans with the method changed, but ratios also differ, there is no support for a significant difference in bitterness, and values are medians of 6 people | |||||
Rune 2026 sensory only
Impact of Water Composition on Coffee Sensory Properties: A Comprehensive Analysis. Journal of Food Science 91(9):e71434 — doi:10.1111/1750-3841.71434
- How deeply checked
- Checked in full text — Full text via Europe PMC (methods, results, Tables 2–6); some table references in the text do not match the tables (the first descriptive analysis is in Table 4) Checked by: LLM (one extracted, another checked against the source; no human check), 2026-10-22.
- Samples
- Colombian Kachalu (same batch, a commercial specialty coffee)
- Processing and cooking
- Described by the producer as a slow-roasted medium roast (no roast-degree value)
- How it was served or eaten
- French press, 50 g grounds to 900 ± 5 g water, 92 °C, 4 min; brewing water made by dissolving calcium, magnesium and bicarbonate salts in ultrapure water; served at about 55 ± 2 °C, 30 mL, for the descriptive analyses
- Who rated, on what scale
- 2 descriptive analyses with 10 trained assessors each (University of Copenhagen); 15 cm line scale (a little to a lot), values in cm; the first with 3 replicates, the second described in 2 places in the methods as 3 and 4 replicates, which disagree; ANOVA and Tukey (p<0.05); separate duo–trio tests with 31 and 25 untrained assessors
| Measured | Comparison and condition | Brewing compared | Comparison | Unit | Direction / test |
|---|---|---|---|---|---|
| Bitterness intensity (sensory) | Water with 300 mg/L bicarbonate added vs ultrapure water Same batch of Colombian Kachalu (medium roast per the producer); French press, 50 g grounds to 900 ± 5 g water, 92 °C, 4 min; water made by dissolving salts in ultrapure water; served at about 55 ± 2 °C, 30 mL | 9.5 (c) | 6.9 (ab) | position on a 15 cm line scale (cm; mean of 10 trained assessors); Tukey letters in brackets (values sharing a letter do not differ significantly) | Brewing compared more bitter Tested, significant |
| Second descriptive analysis; ANOVA over the 5 waters, bitterness p<0.001, Tukey comparison (p<0.05); strength (TDS) and extraction were not measured (an acknowledged limitation); the number of replicates in the second analysis is given in 2 places in the methods as 3 and 4, which disagree; the letters for the 300 mg/L bicarbonate water and ultrapure water do not overlap; sourness in the same table was 2.9 and 12.0; this is the result at this added amount and does not show that any amount makes coffee more bitter Where: Methods; Tables 2 and 5; Results For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Usable — Same batch of beans and the same French press with only water composition changed, rated by 10 trained assessors; strength not measured | |||||
| Bitterness intensity (sensory) | Water with 264 mg/L calcium added vs ultrapure water Same batch of Colombian Kachalu (medium roast per the producer); French press, 50 g grounds to 900 ± 5 g water, 92 °C, 4 min; water made by dissolving salts in ultrapure water; served at about 55 ± 2 °C, 30 mL | 6.0 (a) | 6.9 (ab) | position on a 15 cm line scale (cm; mean of 10 trained assessors); Tukey letters in brackets (values sharing a letter do not differ significantly) | Comparison more bitter Tested, not significant |
| The direction shows which published mean is larger. Second descriptive analysis; ANOVA over the 5 waters, bitterness p<0.001, Tukey comparison (p<0.05); strength (TDS) and extraction were not measured (an acknowledged limitation); the number of replicates in the second analysis is given in 2 places in the methods as 3 and 4, which disagree; the 264 mg/L calcium water and ultrapure water share the letter a, so no significant difference was detected under this calcium condition; this does not show that hardness leaves bitterness unchanged Where: Methods; Tables 2 and 5; Results For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Usable — Same batch of beans and the same French press with only water composition changed, rated by 10 trained assessors; strength not measured | |||||
| Bitterness intensity (sensory) | Water with 264 mg/L calcium and 300 mg/L bicarbonate added vs ultrapure water Same batch of Colombian Kachalu (medium roast per the producer); French press, 50 g grounds to 900 ± 5 g water, 92 °C, 4 min; water made by dissolving salts in ultrapure water; served at about 55 ± 2 °C, 30 mL; water with half of each (132 mg/L calcium and 150 mg/L bicarbonate) was 8.3 (bc), sharing the letter b with ultrapure water | 9.5 (c) | 6.9 (ab) | position on a 15 cm line scale (cm; mean of 10 trained assessors); Tukey letters in brackets (values sharing a letter do not differ significantly) | Brewing compared more bitter Tested, significant |
| Second descriptive analysis; ANOVA over the 5 waters, bitterness p<0.001, Tukey comparison (p<0.05); strength (TDS) and extraction were not measured (an acknowledged limitation); the number of replicates in the second analysis is given in 2 places in the methods as 3 and 4, which disagree; the letters for the water with 264 mg/L calcium and 300 mg/L bicarbonate and ultrapure water do not overlap; bicarbonate added was 300 mg/L, but titration gave 162 mg/L, or 183 mg/L after a 13% correction, which the authors interpret as a loss by precipitation Where: Methods; Tables 2, 5 and 6; Results For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Usable — Same batch of beans and the same French press with only water composition changed, rated by 10 trained assessors; strength not measured | |||||
| Bitterness intensity (sensory) | Water with 132 mg/L calcium and 80 mg/L magnesium added (with or without 60 mg/L bicarbonate) vs ultrapure water Same batch of Colombian Kachalu (medium roast per the producer); French press, 50 g grounds to 900 ± 5 g water, 92 °C, 4 min; water made by dissolving salts in ultrapure water; served at about 55 ± 2 °C, 30 mL; first descriptive analysis (10 trained assessors, 3 replicates) | 10.0 (no bicarbonate), 9.5 (with bicarbonate) | 9.6 | position on a 15 cm line scale (cm; mean of 10 trained assessors) | Depends on the combination Tested, not significant |
| ANOVA over the 3 waters (132 mg/L calcium and 80 mg/L magnesium added, with or without 60 mg/L bicarbonate, and ultrapure water), p=0.864, with no significant difference for any attribute; by published means, the water without bicarbonate is above ultrapure water and the one with bicarbonate below Where: Methods; Tables 2 and 4; Results For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Usable — As above (first descriptive analysis, lower amounts) | |||||
Chen 2023 sensory only
Evaluation of Physicochemical Characteristics and Sensory Properties of Cold Brew Coffees Prepared Using Ultrahigh Pressure under Different Extraction Conditions. Foods 12(20):3857 — doi:10.3390/foods12203857
- How deeply checked
- Checked in full text — Full text via Europe PMC (sections 2.2, 2.8 and 2.9, section 3.3, Tables 1, 4 and 5) Checked by: LLM (one extracted, another checked against the source; no human check), 2026-10-22.
- Samples
- Ethiopian arabica
- Processing and cooking
- Drum-roasted for 11.5 ± 0.5 min to a final 205.0 ± 1.0 °C, medium roast (L value 24–26); ground to 400–600 μm
- How it was served or eaten
- 15 g grounds to 210 g water at 5 °C. Ultra-high pressure: the bottle placed in the unit at 100–500 MPa for 20 min, or at 300 MPa for 10–30 min. Conventional cold brew: atmospheric pressure, 12 h in a refrigerator at 5 °C. All filtered, kept at 5 °C and served at 20 ± 3 °C; strength not equalised
- Who rated, on what scale
- 6 CQI-certified Arabica Q Graders, 2 familiarisation sessions; 0 (none) to 15 (extremely intense) in 0.5 steps, 3 replicates; one-way ANOVA (p<0.05); tables give means ± SD with significance letters
| Measured | Comparison and condition | Brewing compared | Comparison | Unit | Direction / test |
|---|---|---|---|---|---|
| Bitterness intensity (sensory) | Ultra-high pressure: 500 MPa, 20 min vs conventional cold brew (atmospheric, 5 °C, 12 h) The same Ethiopian arabica, roasted medium (L value 24–26) and ground to 400–600 μm; 15 g grounds to 210 g water at 5 °C; for ultra-high pressure the bottle is placed in the unit; conventional cold brew at atmospheric pressure, 12 h in a refrigerator at 5 °C; all filtered, kept at 5 °C and served at 20 ± 3 °C; strength (TDS) 1.22% at 500 MPa and 1.21% for the conventional cold brew | 5.20 ± 0.48 (b) | 6.75 ± 0.53 (a) | 0 (none) to 15 (extremely intense) in 0.5 steps (6 Q Graders, 3 replicates); mean ± SD, significance letters in brackets | Comparison more bitter Tested, significant |
| One-way ANOVA (p<0.05); the letters do not overlap; the conventional cold brew was significantly higher than 300, 400 and 500 MPa; pressure and time (20 min and 12 h) differ together; an industrial process, not a home brewing method Where: Sections 2.2 and 2.8; Tables 1 and 4 For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Limited — Same grounds with the extraction method changed, but ultra-high pressure is an industrial process, not home brewing, and pressure and time differ together | |||||
| Bitterness intensity (sensory) | Ultra-high pressure: 100 MPa vs 500 MPa (both 20 min) The same Ethiopian arabica, roasted medium (L value 24–26) and ground to 400–600 μm; 15 g grounds to 210 g water at 5 °C; for ultra-high pressure the bottle is placed in the unit; conventional cold brew at atmospheric pressure, 12 h in a refrigerator at 5 °C; all filtered, kept at 5 °C and served at 20 ± 3 °C; 200 MPa 5.88 (ab), 300 MPa 5.38 (b), 400 MPa 5.25 (b) | 6.00 ± 0.66 (ab) | 5.20 ± 0.48 (b) | 0 (none) to 15 (extremely intense) in 0.5 steps (6 Q Graders, 3 replicates); mean ± SD, significance letters in brackets | Brewing compared more bitter Tested, not significant |
| The direction shows which published mean is larger; means fell as pressure rose, but no significant difference was shown among the ultra-high-pressure groups (all share the letter b); the abstract's statement that bitterness decreased as pressure rose matches the fall in means but does not say it was significant; the methods give 300–500 MPa, but the table also has 100 and 200 MPa Where: Abstract; section 2.2; Table 4 For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Limited — Same grounds with only pressure changed (time fixed at 20 min), but an industrial process | |||||
| Bitterness intensity (sensory) | Ultra-high pressure: 300 MPa, 30 min vs 10 min The same Ethiopian arabica, roasted medium (L value 24–26) and ground to 400–600 μm; 15 g grounds to 210 g water at 5 °C; for ultra-high pressure the bottle is placed in the unit; conventional cold brew at atmospheric pressure, 12 h in a refrigerator at 5 °C; all filtered, kept at 5 °C and served at 20 ± 3 °C; 15 min 4.85 (bc), 20 min 5.38 (bc), 25 min 5.50 (abc); strength (TDS) 1.11% at 10 min and 1.24% at 30 min | 5.88 ± 0.72 (ab) | 4.25 ± 0.68 (c) | 0 (none) to 15 (extremely intense) in 0.5 steps (6 Q Graders, 3 replicates); mean ± SD, significance letters in brackets | Brewing compared more bitter Tested, significant |
| One-way ANOVA (p<0.05); 10 and 30 min do not share a letter; not every pair of adjacent times differs; 30 min and the conventional cold brew (6.75 (a)) share a letter Where: Tables 1 and 5 For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Limited — Same grounds with pressure fixed at 300 MPa and time varied, but an industrial process | |||||
Shi 2024 sensory only
Comparison of Ultra-High-Pressure and Conventional Cold Brew Coffee at Different Roasting Degrees: Physicochemical Characteristics and Volatile and Non-Volatile Components. Foods 13(19):3119 — doi:10.3390/foods13193119
- How deeply checked
- Checked in full text — Full text via Europe PMC (Tables 1–5); the abstract says the compounds increased, but Table 2 and the text say they decreased Checked by: LLM (one extracted, another checked against the source; no human check), 2026-10-21.
- Samples
- Ethiopian arabica (bought from the same company; whether the same lot, the variety and processing are not given)
- Processing and cooking
- States only that light, medium and dark correspond to Agtron color value zones 73, 62 and 53; the scale type and roasting conditions are not given
- How it was served or eaten
- 15 g grounds to 210 mL water; ultra-high pressure (300 MPa, 20 min) or conventional cold brew (4 °C, 12 h); strength not equalised
- Who rated, on what scale
- 12 Q Arabica graders; rated at 20 ± 3 °C, 3 replicates; scale range not given; means ± SD with significance letters (5% level)
| Measured | Comparison and condition | Brewing compared | Comparison | Unit | Direction / test |
|---|---|---|---|---|---|
| Bitterness intensity (sensory) | Ultra-high pressure: conventional cold brew vs 300 MPa, 20 min (light roast) Ethiopian arabica, light roast (Agtron colour zone 73); 15 g grounds to 210 mL water; ultra-high pressure at 300 MPa for 20 min (water temperature not given) and conventional cold brew at 4 °C for 12 h; both filtered, kept at 4 °C and served at 20 ± 3 °C; strength (TDS) 1.11% for the conventional cold brew and 1.10% under ultra-high pressure (no significant difference within the same roast) | 3.00 ± 0.28 (d) | 2.00 ± 0.10 (e) | scale range not given (12 Q Arabica Graders); mean ± SD, significance letters in brackets | Brewing compared more bitter Tested, significant |
| ANOVA (p<0.05); the letters do not overlap; pressure and time (20 min and 12 h) differ together; the authors overlap with Chen et al., so this is not counted as an independent replication Where: Sections 2.1 and 2.7; Tables 1 and 5; section 3.5 For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Limited — Same beans and roast with the extraction method changed, but ultra-high pressure is an industrial process, pressure and time differ together, and the scale range is not given | |||||
| Bitterness intensity (sensory) | Ultra-high pressure: conventional cold brew vs 300 MPa, 20 min (medium roast) Ethiopian arabica, medium roast (Agtron colour zone 62); 15 g grounds to 210 mL water; ultra-high pressure at 300 MPa for 20 min (water temperature not given) and conventional cold brew at 4 °C for 12 h; both filtered, kept at 4 °C and served at 20 ± 3 °C; strength (TDS) 1.21% for the conventional cold brew and 1.18% under ultra-high pressure (no significant difference within the same roast) | 4.30 ± 0.42 (c) | 3.35 ± 0.30 (d) | scale range not given (12 Q Arabica Graders); mean ± SD, significance letters in brackets | Brewing compared more bitter Tested, significant |
| ANOVA (p<0.05); the letters do not overlap; pressure and time (20 min and 12 h) differ together; the authors overlap with Chen et al., so this is not counted as an independent replication Where: Sections 2.1 and 2.7; Tables 1 and 5; section 3.5 For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Limited — Same beans and roast with the extraction method changed, but ultra-high pressure is an industrial process, pressure and time differ together, and the scale range is not given | |||||
| Bitterness intensity (sensory) | Ultra-high pressure: conventional cold brew vs 300 MPa, 20 min (dark roast) Ethiopian arabica, dark roast (Agtron colour zone 53); 15 g grounds to 210 mL water; ultra-high pressure at 300 MPa for 20 min (water temperature not given) and conventional cold brew at 4 °C for 12 h; both filtered, kept at 4 °C and served at 20 ± 3 °C; strength (TDS) 1.30% for the conventional cold brew and 1.28% under ultra-high pressure (no significant difference within the same roast) | 6.50 ± 0.42 (a) | 5.50 ± 0.40 (b) | scale range not given (12 Q Arabica Graders); mean ± SD, significance letters in brackets | Brewing compared more bitter Tested, significant |
| ANOVA (p<0.05); the letters do not overlap; pressure and time (20 min and 12 h) differ together; the authors overlap with Chen et al., so this is not counted as an independent replication Where: Sections 2.1 and 2.7; Tables 1 and 5; section 3.5 For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Limited — Same beans and roast with the extraction method changed, but ultra-high pressure is an industrial process, pressure and time differ together, and the scale range is not given | |||||
Córdoba 2019 sensory only
Effect of grinding, extraction time and type of coffee on the physicochemical and flavour characteristics of cold brew coffee. Scientific Reports 9:8440 — doi:10.1038/s41598-019-44886-w
- How deeply checked
- Checked in full text — Full text via Europe PMC (methods, the Flavour profile results, Tables 2 and 3) Checked by: LLM (one extracted, another checked against the source; no human check), 2026-10-22.
- Samples
- Colombian specialty arabica (Caturra) from 2 origins, Huila and Nariño
- Processing and cooking
- Medium roast (no numerical roasting conditions in the parts read)
- How it was served or eaten
- Cold: immersion at 20 °C, 60 g grounds to 700 g water, coarse (701–900 μm) or medium (501–700 μm) grind for 14 or 22 h. Hot: French press with boiling water, 42.5 g grounds to 500 g water. Cold brews served at 19 °C and hot brews at 70 °C, in separate sessions
- Who rated, on what scale
- 3 experienced tasters (over 3 years of experience, over 40 h of training); 0 (not perceptible) to 5 (very strong), 3 replicates per sample, 2 blocks; one-way ANOVA and LSD (5%); tables give means ± 95% confidence intervals (n=6)
| Measured | Comparison and condition | Brewing compared | Comparison | Unit | Direction / test |
|---|---|---|---|---|---|
| Bitterness intensity (sensory) | Coarse vs medium grind (Huila, 14 h cold brew) Colombian Caturra (medium roast); cold brew by immersion at 20 °C, 60 g grounds to 700 g water, served at 19 °C; other pairs: Nariño 14 h 2.17 (a) and 2.00 (ab), Huila 22 h 1.67 (ab) and 1.17 (ab), Nariño 22 h 2.17 (a) and 1.50 (ab) | 1.50 (ab) | 1.00 (b) | taste bitterness, 0 (not perceptible) to 5 (very strong) (3 experienced tasters, 3 replicates per sample; table gives means ± 95% confidence intervals, n=6); LSD letters in brackets | Brewing compared more bitter Tested, not significant |
| The direction shows which published mean is larger; one-way ANOVA and LSD (5%); for grind within the same origin and time, every pair shares a letter, so no significant difference was detected Where: Methods; Table 3 (Taste, Bitter) For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Limited — Cold brews of the same origin with only grind or time changed, but 3 people rated | |||||
| Bitterness intensity (sensory) | 22 h vs 14 h (Huila, coarse-ground cold brew) Colombian Caturra (medium roast); cold brew by immersion at 20 °C, 60 g grounds to 700 g water, served at 19 °C; other pairs: Nariño coarse 2.17 (a) and 2.17 (a), Huila medium 1.17 (ab) and 1.00 (b), Nariño medium 1.50 (ab) and 2.00 (ab) | 1.67 (ab) | 1.50 (ab) | taste bitterness, 0 (not perceptible) to 5 (very strong) (3 experienced tasters, 3 replicates per sample; table gives means ± 95% confidence intervals, n=6); LSD letters in brackets | Brewing compared more bitter Tested, not significant |
| The direction shows which published mean is larger; one-way ANOVA and LSD (5%); for time within the same origin and grind, every pair shares a letter, so no significant difference was detected; the direction of the means also differs between pairs Where: Methods; Table 3 (Taste, Bitter) For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Limited — Cold brews of the same origin with only grind or time changed, but 3 people rated | |||||
| Bitterness intensity (sensory) | Reference (serving temperature and session differ): hot brew (French press) vs 4 cold brews (Huila) Colombian Caturra (medium roast); hot: French press with boiling water, 42.5 g grounds to 500 g water, served at 70 °C; cold: 20 °C, 60 g grounds to 700 g water, 4 versions of coarse or medium grind × 14 or 22 h, served at 19 °C; hot and cold brews rated in separate sessions | 2.00 (ab) | 1.00–1.67 (b to ab) | taste bitterness, 0 (not perceptible) to 5 (very strong) (3 experienced tasters, 3 replicates per sample; table gives means ± 95% confidence intervals, n=6); LSD letters in brackets | Brewing compared more bitter Tested, not significant |
| The direction shows which published mean is larger. No significant difference in taste bitterness is shown between the hot brew and each cold brew of the same origin (letters overlap); serving temperature and session differ, so this is not a comparison of brewing alone; the text sums up the aftertaste bitterness of cold brews as higher, but the table shows no significant difference in aftertaste bitterness within the same origin either; aftertaste bitterness was 2.00 (ab) hot and 1.33–2.33 cold Where: Methods, "Sensory evaluation"; Results, "Flavour profile"; Table 3 For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Not usable — Hot and cold brews were served at different temperatures (70 °C and 19 °C) and rated in different sessions, so this cannot be read as a comparison of brewing alone; 3 people rated | |||||
| Bitterness intensity (sensory) | Reference (serving temperature and session differ): hot brew (French press) vs 4 cold brews (Nariño) Colombian Caturra (medium roast); hot: French press with boiling water, 42.5 g grounds to 500 g water, served at 70 °C; cold: 20 °C, 60 g grounds to 700 g water, 4 versions of coarse or medium grind × 14 or 22 h, served at 19 °C; hot and cold brews rated in separate sessions | 2.00 (ab) | 1.50–2.17 (ab to a) | taste bitterness, 0 (not perceptible) to 5 (very strong) (3 experienced tasters, 3 replicates per sample; table gives means ± 95% confidence intervals, n=6); LSD letters in brackets | Depends on the combination Tested, not significant |
| By published means, 2 of the cold brews are above the hot brew, 1 equal and 1 below. No significant difference in taste bitterness is shown between the hot brew and each cold brew of the same origin (letters overlap); serving temperature and session differ, so this is not a comparison of brewing alone; the text sums up the aftertaste bitterness of cold brews as higher, but the table shows no significant difference in aftertaste bitterness within the same origin either; aftertaste bitterness was 2.50 (a) hot and 1.67–2.67 cold Where: Methods, "Sensory evaluation"; Results, "Flavour profile"; Table 3 For whether changing only how the same beans (same origin and roast) are brewed (cold or hot water, water temperature, equipment, extraction pressure or time, water composition, strength) changes the intensity of bitterness people perceive: Not usable — Hot and cold brews were served at different temperatures (70 °C and 19 °C) and rated in different sessions, so this cannot be read as a comparison of brewing alone; 3 people rated | |||||
What was established within an experiment
Only results observed in comparisons within one study. Results with no significant difference detected are stated as such.
- For the same beans in Cai et al., diluted to 2% strength and served at 20 ± 3 °C, bitterness was 4.14 ± 0.69 hot and 2.43 ± 0.53 cold, significantly higher hot (t-test, p<0.05). Astringency was also higher hot (2.86 and 2.00), while sourness was 2.86 and 3.29 with no significant difference detected. [rows: Cai 2022 Hot brew (coffee machine) vs cold brew (8 h at room temperature)]
- For the 4 pairs of the same beans in Heo et al., bitterness intensity (9-point scale) for coffee maker and cold brew was 4.9 and 4.9 for the blend and 4.9 and 4.6 for Colombia, with no significant difference detected, while for Folgers (6.8 and 5.8) and Kenya (5.0 and 4.4) the coffee maker brew was significantly higher. The raters were 120 untrained consumers, and this item is separate from bitterness liking. [rows: Heo 2019 Coffee maker vs cold brew (blend); Heo 2019 Coffee maker vs cold brew (Colombia); Heo 2019 Coffee maker vs cold brew (Folgers); Heo 2019 Coffee maker vs cold brew (Kenya)]
- In the immersions of Batali et al., in the main effect of temperature pooled over 9 coffees, 92 °C was significantly more bitter than 4 °C and 22 °C, with no significant difference detected between 4 °C and 22 °C (F=6.06). In Liang et al., in results pooling the 5 time points for each roast, bitterness increased significantly from 4 °C < 22 °C < 92 °C for the darker roast, while for the lighter roast no significant difference was detected between 4 °C and 22 °C, both lower than 92 °C. In both studies the values are in figures, which were not obtained, so the numbers are unconfirmed. [rows: Batali 2022 Overall result (pooled over origins and roasts): 92 °C vs 4 °C immersion (9 coffees); Batali 2022 Overall result (pooled over origins and roasts): 22 °C vs 4 °C immersion (9 coffees); Liang 2024 92 °C and 22 °C vs 4 °C immersion (darker roast); Liang 2024 92 °C vs 4 °C and 22 °C immersion (lighter roast); Liang 2024 22 °C vs 4 °C immersion (lighter roast)]
- For the same beans in Lopane et al., the text says the hot drip (about 97 °C, 1:18, flash-chilled) was significantly more bitter than the 4 °C and 25 °C immersions (1:14) on both day 0 and day 42 of storage (served at 7 °C and rated 0–15 by 5 professionals). The coffee-to-water ratio also differs, so this is not a comparison of temperature alone. [rows: Lopane 2024 Hot drip vs 4 °C immersion (storage day 0); Lopane 2024 Hot drip vs 4 °C immersion (storage day 42)]
- In the drip brews of Batali et al. 2020, with strength and extraction matched and brew temperature varied over 87, 90 and 93 °C, bitterness was not among the attributes that varied significantly with temperature. Over strengths (TDS) from about 1% to about 1.5%, bitterness increased with strength. For the paper drip of Izumi and Takaya 2008, no significant difference was detected among the 3 temperatures of 85, 90 and 95 °C for attributes other than colour. [rows: Batali 2020 Drip at 93 °C vs 87 °C (and 90 °C), strength and extraction matched; Batali 2020 Strength (TDS) about 1.5% vs about 1% (pooled over temperature and extraction); Izumi and Takaya 2008 Paper drip at 95 °C vs 85 °C (and 90 °C) (3 coffees)]
- In the French press brews of the same beans in Rune et al., bitterness with 300 mg/L bicarbonate added was 9.5 against 6.9 for ultrapure water, significantly higher, and water with both 264 mg/L calcium and 300 mg/L bicarbonate added was also 9.5. With only 264 mg/L calcium it was 6.0, with no significant difference from ultrapure water detected. At lower amounts (132 mg/L calcium and 80 mg/L magnesium, with or without 60 mg/L bicarbonate) values were 10.0 and 9.5 against 9.6 for ultrapure water, with no significant difference detected (p=0.864). [rows: Rune 2026 Water with 300 mg/L bicarbonate added vs ultrapure water; Rune 2026 Water with 264 mg/L calcium added vs ultrapure water; Rune 2026 Water with 264 mg/L calcium and 300 mg/L bicarbonate added vs ultrapure water; Rune 2026 Water with 132 mg/L calcium and 80 mg/L magnesium added (with or without 60 mg/L bicarbonate) vs ultrapure water]
- For coffee A in Izumi and Takaya, the siphon was about 3.3 and paper drip about 2.3 (1–5 points, read by eye from Fig. 1), and for coffee B hard water was about 3.7 and distilled water about 2.8 (Fig. 2); in both, bitterness was marked p<0.01 in the comparison of 3 conditions. Which 2 conditions differ is not shown. [rows: Izumi and Takaya 2008 Siphon vs paper drip (coffee A; grind also differs); Izumi and Takaya 2008 Commercial hard water vs distilled water (coffee B; a comparison of the commercial waters used)]
- For the same grounds in Chen et al., bitterness of 5.20 at 500 MPa for 20 min was significantly lower than 6.75 for conventional cold brew (5 °C, 12 h). From 100 MPa to 500 MPa the mean fell from 6.00 to 5.20, but no significant difference was shown among the ultra-high-pressure groups. At 300 MPa, extending the holding time from 10 to 30 min raised bitterness significantly from 4.25 to 5.88. In Shi et al., for light, medium and dark roasts alike, conventional cold brew (3.00, 4.30, 6.50) was significantly higher than ultra-high pressure (2.00, 3.35, 5.50). [rows: Chen 2023 Ultra-high pressure: 500 MPa, 20 min vs conventional cold brew (atmospheric, 5 °C, 12 h); Chen 2023 Ultra-high pressure: 100 MPa vs 500 MPa (both 20 min); Chen 2023 Ultra-high pressure: 300 MPa, 30 min vs 10 min; Shi 2024 Ultra-high pressure: conventional cold brew vs 300 MPa, 20 min (light roast); Shi 2024 Ultra-high pressure: conventional cold brew vs 300 MPa, 20 min (medium roast); Shi 2024 Ultra-high pressure: conventional cold brew vs 300 MPa, 20 min (dark roast)]
- For the same beans in Vezzulli et al., median bitterness for arabica was 4.0 for espresso, 3.4 for moka, 3.1 for Neapolitan pot and 1.2 for filter, and for robusta 4.7 for espresso and 3.5 for filter (0–9). ANOVA was carried out, but no result of the bitterness comparison is reported (bitterness is not in the list of significant attributes). [rows: Vezzulli 2022 Espresso vs filter (arabica); Vezzulli 2022 Espresso vs filter (robusta)]
- In the cold brews of Córdoba et al., changing grind (coarse or medium) and time (14 or 22 h) within the same origin produced no significant difference detected in taste bitterness (0–5). The comparison with the hot brew (French press) cannot be read as a comparison of brewing alone, because serving temperature (70 °C and 19 °C) and session differ. [rows: Córdoba 2019 Coarse vs medium grind (Huila, 14 h cold brew); Córdoba 2019 22 h vs 14 h (Huila, coarse-ground cold brew); Córdoba 2019 Reference (serving temperature and session differ): hot brew (French press) vs 4 cold brews (Huila); Córdoba 2019 Reference (serving temperature and session differ): hot brew (French press) vs 4 cold brews (Nariño)]
Candidate explanations for the differences
Conditions that differ between studies and could explain the differences in bitterness results, and the authors' interpretations. Several conditions differ at once, so which one mattered is unknown. These are candidates, not causes.
- A candidate explanation for differences between studies is how strength and serving temperature were matched. Cai et al. and Batali et al. 2022, where the hot brew was significantly more bitter, diluted to 2% strength and served at one temperature. Heo et al. brewed at the same coffee-to-water ratio (7 g grounds to 100 mL water) with no record of matching strength, served both at about 8 °C, and detected no significant difference in 2 of the 4 pairs. However, beans, scales and raters also differ at the same time, so which of these mattered has not been confirmed. [rows: Cai 2022 Hot brew (coffee machine) vs cold brew (8 h at room temperature); Batali 2022 Overall result (pooled over origins and roasts): 92 °C vs 4 °C immersion (9 coffees); Heo 2019 Coffee maker vs cold brew (blend); Heo 2019 Coffee maker vs cold brew (Colombia); Heo 2019 Coffee maker vs cold brew (Folgers); Heo 2019 Coffee maker vs cold brew (Kenya)]
- Strength (TDS) is another candidate. In Batali et al. 2020 bitterness rose with strength even at matched temperatures, and Liang et al. 2024 report that bitterness correlated positively with strength. In Lopane et al., however, the 25 °C immersion, which had the highest strength, was less bitter than the hot drip, so strength alone does not predict bitterness in these results. Rune et al. did not measure strength. [rows: Batali 2020 Strength (TDS) about 1.5% vs about 1% (pooled over temperature and extraction); Liang 2024 92 °C and 22 °C vs 4 °C immersion (darker roast); Lopane 2024 Hot drip vs 4 °C immersion (storage day 0); Rune 2026 Water with 300 mg/L bicarbonate added vs ultrapure water]
- On the difference between ultra-high pressure and conventional cold brew, Chen et al. argue that holding time affects trigonelline and chlorogenic acid and influences bitterness more strongly than pressure. This is the authors' interpretation; ultra-high pressure and conventional cold brew differ in pressure and time (20 min and 12 h) at once, so which mattered cannot be separated. [rows: Chen 2023 Ultra-high pressure: 500 MPa, 20 min vs conventional cold brew (atmospheric, 5 °C, 12 h); Chen 2023 Ultra-high pressure: 300 MPa, 30 min vs 10 min; Shi 2024 Ultra-high pressure: conventional cold brew vs 300 MPa, 20 min (medium roast)]
What is not known — what this comparison alone cannot separate
- Scales and raters differ by study (1–5 points, 0–5, a 9-point scale, medians on 0–9, 0–15, a 15 cm line, 0–100, and one with no scale range given; from 3 experienced tasters to 120 untrained consumers). Values are not compared across studies. [rows: Cai 2022 Hot brew (coffee machine) vs cold brew (8 h at room temperature); Córdoba 2019 Coarse vs medium grind (Huila, 14 h cold brew); Heo 2019 Coffee maker vs cold brew (blend); Vezzulli 2022 Espresso vs filter (arabica); Lopane 2024 Hot drip vs 4 °C immersion (storage day 0); Rune 2026 Water with 300 mg/L bicarbonate added vs ultrapure water; Batali 2022 Overall result (pooled over origins and roasts): 92 °C vs 4 °C immersion (9 coffees); Shi 2024 Ultra-high pressure: conventional cold brew vs 300 MPa, 20 min (light roast)]
- Temperature ranges are not merged across studies. The studies that detected no significant difference across hot brew temperatures (87–93 °C and 85–95 °C) and those comparing immersions at 4, 22 and 92 °C differ in brewing, strength and serving temperature. No row compares the same beans at temperatures stepped between these. [rows: Batali 2020 Drip at 93 °C vs 87 °C (and 90 °C), strength and extraction matched; Izumi and Takaya 2008 Paper drip at 95 °C vs 85 °C (and 90 °C) (3 coffees); Batali 2022 Overall result (pooled over origins and roasts): 92 °C vs 4 °C immersion (9 coffees); Liang 2024 92 °C and 22 °C vs 4 °C immersion (darker roast)]
- No study that varied espresso pump pressure on the same beans with people rating bitterness was found in the search record of the reading step (a record of where we looked, not proof that none exists). In the equipment comparisons, siphon and paper differ also in grind and in the amounts of coffee and water (the ratio is 1:15 for both), and espresso and filter differ also in extraction style and coffee-to-water ratio, so the effect of the device itself cannot be separated. The water rows are only comparisons of salts added to ultrapure water and of commercial waters whose composition is not given; no row steps through the range of tap-water hardness. [rows: Izumi and Takaya 2008 Siphon vs paper drip (coffee A; grind also differs); Vezzulli 2022 Espresso vs filter (arabica); Rune 2026 Water with 300 mg/L bicarbonate added vs ultrapure water; Rune 2026 Water with 264 mg/L calcium added vs ultrapure water; Izumi and Takaya 2008 Commercial hard water vs distilled water (coffee B; a comparison of the commercial waters used)]
- Lopane et al. write in the same text in ways that can be read as bitterness rising and as falling with storage, and the figures were not obtained, so the change with storage is not used on this page. In comparisons where chilled holding times differ (cold brew and coffee maker in Heo et al.), the effect of holding time also cannot be separated. [rows: Lopane 2024 Hot drip vs 4 °C immersion (storage day 42); Heo 2019 Coffee maker vs cold brew (Kenya)]
How studies and rows were counted
The number of evidence rows is not the number of studies. Table rows, ledger evidence rows, papers and research groups are counted separately.
- Sources in the table: 12, of which 12 have results on people's bitterness ratings in this table; for the others, only results on something else (compounds, smell) are shown, or they are used only for context.
- Rows in the table (study × measured result): 33. Rows on perceived bitterness: 33.
- Evidence rows in the lab's ledger these rows rest on: 0. One paper can supply several rows. Rows without a ledger evidence row come from other tables of the same papers, or from sources newly read for this question.
- Research groups among the studies with bitterness results in this table, as grouped in the ledger: 9 (reports from the same lab count as 1; co-authors and affiliations can overlap, so this is not a count of independent replications).
What we looked at
The sources examined for this question and why each was included or excluded. We searched Europe PMC and J-STAGE for studies combining coffee brewing (cold brew, water temperature, equipment, water, pressure, grind, time) with human ratings of bitterness, and added the sources already in the lab's ledger. Studies with no human bitterness rating, no item for bitterness intensity alone, or no comparison of brewing were excluded. Studies whose values are only in figures or supplementary tables we could not obtain are marked as such. Studies seen only as citations, whose titles we have not confirmed in an authoritative record, are not listed. The search described here is a record, not a claim that nothing was missed.
| Source | Decision | Why |
|---|---|---|
| Comparative Profiling of Hot and Cold Brew Coffee Flavor Using Chromatographic and Sensory Approaches (2022) doi | Included in the table | The same batch of beans brewed in a coffee machine and as room-temperature cold brew, with strength and serving temperature matched and 7 trained panellists rating; read in full |
| Cold Brew Coffee: Consumer Acceptability and Characterization Using the Check-All-That-Apply (CATA) Method (2019) doi | Included in the table | Mainly a consumer study of commercial cold brews, but includes 4 sets of the same beans brewed cold and in a coffee maker; bitterness intensity (a column separate from liking) is tabled; read in full |
| An investigation of the shelf life of cold brew coffee and the influence of extraction temperature using chemical, microbial, and sensory analysis (2024) doi | Included in the table | The same beans brewed as hot drip and immersions at 2 temperatures, rated by 5 professionals; values are only in figures, which were not obtained, so they are unconfirmed; only the direction and tests in the text are used; read in full |
| Sensory Analysis of Full Immersion Coffee: Cold Brew Is More Floral, and Less Bitter, Sour, and Rubbery Than Hot Brew (2022) doi | Included in the table | 9 coffees each immersed at 3 temperatures, rated with strength and serving temperature matched; the available result is the temperature main effect pooled over origins and roasts, so it is tabled as an overall result; the figure with the bitterness means was not obtained; read in full |
| Sensory analysis of the flavor profile of full immersion hot, room temperature, and cold brewed coffee over time (2024) doi | Included in the table | Two roasts of the same blend immersed at different temperatures and times, rated by 14 trained panellists; temperature results by roast are used as comparisons pooling 5 time points; Fig. 5 is not in the store for this question (the note in the shared ledger that Fig. 5 values were read by eye is a record from the dark-roast question; the bitterness values by temperature for this question were not checked); read in full |
| Brew temperature, at fixed brew strength and extraction, has little impact on the sensory profile of drip brew coffee (2020) doi | Included in the table | Drip brews of the same beans with strength and extraction matched and brew temperature varied, rated by 12 trained panellists; a row for the effect of strength is also tabled; the figures with means and the supplementary tables were not obtained; read in full |
| コーヒーの味に及ぼす抽出条件およびクロロゲン酸量の影響 (2008) doi | Included in the table | A Japanese study varying water temperature, equipment and water for 3 commercial coffees, rated by 12–14 women; figure values were read by eye from page images; the water comparison is of the commercial waters used and is not treated as an effect of hardness alone; read in full (PDF) |
| Metabolomics Combined with Sensory Analysis Reveals the Impact of Different Extraction Methods on Coffee Beverages from Coffea arabica and Coffea canephora var. Robusta (2022) doi | Included in the table | The same beans brewed 4 ways (espresso, moka, Neapolitan pot, filter) and rated by 6 trained panellists; arabica versus robusta belongs to another page, and only brewing comparisons within each coffee are used here; no result of the bitterness comparison is reported; read in full |
| Impact of Water Composition on Coffee Sensory Properties: A Comprehensive Analysis (2026) doi | Included in the table | The same batch of beans brewed in the same French press with water containing added calcium, magnesium and bicarbonate, rated by 10 trained assessors in each analysis; read in full |
| Evaluation of Physicochemical Characteristics and Sensory Properties of Cold Brew Coffees Prepared Using Ultrahigh Pressure under Different Extraction Conditions (2023) doi | Included in the table | The same grounds with ultra-high pressure and time varied and compared with conventional cold brew, rated by 6 Q Graders; an industrial process, so its labels in the list begin with "Ultra-high pressure"; read in full |
| Comparison of Ultra-High-Pressure and Conventional Cold Brew Coffee at Different Roasting Degrees: Physicochemical Characteristics and Volatile and Non-Volatile Components (2024) doi | Included in the table | Ultra-high pressure and conventional cold brew compared at each of 3 roasts of the same beans; the effect of roast belongs to the dark-roast page, and only comparisons within the same roast are used here; authors overlap with Chen et al.; read in full |
| Effect of grinding, extraction time and type of coffee on the physicochemical and flavour characteristics of cold brew coffee (2019) doi | Included in the table | Comparisons of grind and time among cold brews of the same origin are used; rows comparing with the hot brew are kept, but because serving temperature and session differ they cannot be used as main comparisons (marked as not comparable); read in full |
| Coffee brewing sonoreactor for reducing the time of cold brew from several hours to minutes while maintaining sensory attributes (2024) doi | Read as context (not in the table) | The same medium-roast blend extracted for minutes with ultrasound and as 24 h cold brew at 4 °C, rated 0–100 by 11 trained panellists (a comparison of whole methods that also differ in grind, time and temperature); the text says the 3 min ultrasonic brew was more bitter, but the values are only in Fig. 6 and a supplementary table, which were not obtained, so values and pairwise tests are unconfirmed |
| Between Bitterness and Sweetness: How Decaffeination and Sweeteners Shape the Sensory Experience of Espresso Coffee (2026) doi | Excluded | Espresso extraction was fixed, comparing decaffeinated beans and sweeteners; not a comparison of brewing |
| Optimization of Sensory Properties of Cold Brew Coffee Produced by Reduced Pressure Cycles and Its Physicochemical Characteristics (2022) doi | Excluded | Cold brew with repeated reduced-pressure cycles; the sensory evaluation has 6 items such as aroma and no item for bitterness intensity alone |
| Comparative Decoding of Physicochemical and Flavor Profiles of Coffee Prepared by High-Pressure Carbon Dioxide, Ice Drip, and Traditional Cold Brew (2025) doi | Excluded | Human evaluation covers only 8 aromas smelled by nose, with no bitterness rating |
| Effects of ultrasound treatment on physicochemical, sensory, and structural properties of cold brew coffee (2025) doi | Excluded | The sensory evaluation includes sweetness, sourness, nutty, burnt and body, but no item for bitterness intensity |
| Feasibility of Ultrasound-Assisted Extraction for Accelerated Cold Brew Coffee Processing: Characterization and Comparison With Conventional Brewing Methods (2022) doi | Excluded | Sensory evaluation is left for future work, so there is no human bitterness rating |
| Effects of High Pressure Processing and Ultrasound-Assisted Extraction on Physicochemical Properties, Antioxidant Activity and Flavor Compounds of Cold Brew Citri Reticulatae Pericarpium Beverage (2026) doi | Excluded | A tangerine-peel beverage, not coffee |
| Effect of Extraction Methods on Aroma Profile, Antioxidant Activity and Sensory Acceptability of Specialty Coffee Brews (2023) doi | Excluded | Compares extraction methods, but the sensory evaluation has no item for bitterness intensity alone |
| Discrimination of Filter Coffee Extraction Methods of a Medium Roasted Specialty Coffee Based on Volatile Profiles and Sensorial Traits (2023) doi | Excluded | Compares filter methods, but the sensory evaluation has no item for bitterness intensity alone |
| Influence of coffee brewing methods on the chromatographic and spectroscopic profiles, antioxidant and sensory properties (2021) doi | Excluded | Compares brewing methods, but the sensory evaluation has no item for bitterness intensity alone |
| Optimization of Cold Brew Coffee Using Central Composite Design and Its Properties Compared with Hot Brew Coffee (2023) doi | Excluded | Instrumental and compositional measurements only, with no human bitterness rating |
| Cold Brew Coffee—Pilot Studies on Definition, Extraction, Consumer Preference, Chemical Characterization and Microbiological Hazards (2021) doi | Excluded | Includes extraction comparisons with the same arabica and canephora beans, but measured preference ranking and discrimination, not bitterness intensity directly |
| Impact of beverage temperature on consumer preferences for black coffee (2022) doi | Excluded | Mainly analyses preference by drinking temperature; bitterness intensity by brewing method cannot be taken from this report |
| Cross-Cultural Comparison of the Espresso Protocol Repeatability (2025) doi | Excluded | About the repeatability of an espresso evaluation protocol, with beans and roasts not matched between countries; not a comparison of brewing |
Read in the encyclopedia
- Coffee brew — a drink whose bitterness is reshaped not just by caffeine but by roasting byproducts and brewing conditions
- World Coffee Research Sensory Lexicon — a lexicon that lets trained panels rate intensities such as bitterness on a common footing
- Coffee bitterness — the topic hub on coffee bitterness
- Is darker-roast coffee more bitter? — a comparison of roast and bitterness
- Arabica or robusta: which is more bitter? — a comparison of species and bitterness
Source ledger
Each study's source can be reached from its section above. All of the lab's sources are listed in the source ledger.
How to cite this page
Aoki, M. (NIGAMI LAB) (2026). Does brewing method change coffee's bitterness?. https://nigamilab.com/en/topics/coffee-brewing-bitterness/
Each claim card has a stable anchor (#clm-…). The dataset is CC BY 4.0; see the data page. Dataset DOI: 10.5281/zenodo.22739101