Water minerals and flavour
What studies have measured about how the minerals in brewing water change coffee – alkalinity lowers perceived acidity, while the effects of hardness, magnesium and sodium are smaller or unsettled.
Across the chain: Science
Contents
In brief
- Alkalinity (mostly bicarbonate) neutralises part of the extracted coffee acids; the higher it is, the lower the perceived acidity.1,3,10
- In a 2026 sensory study, adding bicarbonate reduced acidity, fruitiness and complexity and raised roasted and bitter notes; hardness alone produced no marked change.10
- That harder water extracts more rests mainly on a computer model; laboratory tests reported by the SCA found no difference measurable with a refractometer.1,5
- Single ions at their taste thresholds in water did not change the flavour of brewed coffee in a 1955 study.7
- Much bicarbonate also slows the flow of water through the coffee bed, which lengthens contact time.1,2,3
What water can change in the cup#
Brewing water does more than dissolve coffee. Its minerals can react with what is extracted, change how fast water passes through the grounds and add flavours of their own.1,2 The water for coffee entry explains where these minerals come from and which targets are published. This entry looks at the evidence for their effects.
Two numbers describe most waters: total hardness (calcium plus magnesium) and alkalinity (the capacity to neutralise acid, mostly bicarbonate).1 The hardness and alkalinity entry defines both. Other ions, such as sodium, chloride and sulphate, matter mainly at high levels.2,7
The evidence is uneven. Some effects have been measured in brewed coffee with tasting panels. Others come from a computer model, from correlations across bottled waters or from handbook summaries of older tests. This entry keeps them apart.
Alkalinity and acidity#
Alkalinity has the clearest effect. Bicarbonate in the water reacts with the acids extracted from the coffee and neutralises part of them.3 The higher the alkalinity, the lower the perceived acidity.1
Tasting studies agree on the direction. In older tests summarised by the SCAA handbook, coffee brewed with water containing carbonates was the most bitter and tasted flat.2 In a 2026 sensory study, adding bicarbonate reduced perceived acidity, fruitiness and complexity.10 Roasted and bitter notes rose at the same time.10
Neutralising an acid with bicarbonate releases carbon dioxide.1 The acids in coffee are covered in their own entry.
Hardness#
Hardness is often said to raise extraction. The idea comes mainly from a 2014 computer model of how calcium, magnesium and sodium ions bind coffee compounds.5 It was a calculation, not a brewing trial.5
Measurements are less clear. In laboratory tests reported by the Specialty Coffee Association (SCA), a coffee refractometer showed no effect of hardness on extraction.1 The same tests found a clear effect on flavour balance and aroma.1 In the 2026 sensory study, hardness alone produced no marked flavour change, even with calcium above usual drinking-water levels.10
Studies with natural or bottled waters cannot always separate hardness from alkalinity. In a 2021 filter-coffee study, sourness fell and bitterness rose from the softest to the hardest of four waters.9 That study measured hardness and dissolved solids, not alkalinity.9 In most natural waters, hardness and alkalinity are roughly equal, so they tend to rise and fall together.1
Data table
| Point | Intensity, 0–5 scale (trained panel of 10) | Source | |
|---|---|---|---|
| Purified water (9 ppm) | bitter 1.30 | 1.3 | 9 |
| Purified water (9 ppm) | sour 3.35 | 3.35 | 9 |
| Bottled water (37 ppm) | bitter 1.55 | 1.55 | 9 |
| Bottled water (37 ppm) | sour 2.88 | 2.88 | 9 |
| Medium-hard mains (72 ppm) | sour 2.55 | 2.55 | 9 |
| Medium-hard mains (72 ppm) | bitter 2.82 | 2.82 | 9 |
| Hard mains (158 ppm) | sour 1.52 | 1.52 | 9 |
| Hard mains (158 ppm) | bitter 3.63 | 3.63 | 9 |
Whether calcium and magnesium act differently is not settled. The SCA cites theory and preliminary experiments suggesting that magnesium extracts slightly more; the effect on flavour remains unclear.1
Sodium, chloride and other ions#
Single ions at low levels had little effect in an early study. In 1955, coffee brewed with water holding one ion at its taste threshold could not be told apart from coffee made with distilled water.7 Concentrations had to be raised further before panels noticed a difference.7
High levels still change the cup. According to the SCAA handbook, much chloride and sulphate can make water taste salty and unbalance the brew.2 It also describes sodium adding sweetness at low levels and strengthening sourness at higher ones; no study opened for this entry tested this.2 In a 2024 study of cold brew made with six commercial waters, more sodium and potassium went with stronger caramel flavours and bitterness.11
Iron is visible before it is tasted. In the 1955 study, iron turned coffee greenish, especially with cream.7
Flow and contact time#
Minerals can also act on the coffee bed. More than any other compounds in tap water, carbonates and bicarbonates slow the flow of water through the grounds, according to the SCAA handbook.2 The SCA handbook explains this with the carbon dioxide formed when bicarbonate neutralises coffee acids.1 A longer contact time can lead to over-extraction.1
Softened water is a special case. Softening swaps calcium and magnesium for sodium but leaves the bicarbonate.2 The SCAA handbook reports that such water slowed brewing and advises against it.2 The water treatment entry covers softeners.
Filter and espresso#
Filter coffee uses far more water per gram of coffee than espresso. The same water therefore brings more buffer to the same acids, and alkalinity matters more in filter.1,3 The water for coffee entry works through the numbers.
Espresso has its own water questions. A review of espresso water discusses alkalinity and individual cations.6 According to a coffee science textbook, hardness has been shown to reduce crema stability.4
Reference values#
| Parameter | Value reported | Conditions | Source |
|---|---|---|---|
| Bicarbonate taste threshold in water (as NaHCO₃) | 770 ppm HCO₃⁻; 290 ppm Na⁺ | Triangle tests, experienced panels | Lockhart et al.7 |
| Bicarbonate detectable in coffee (as NaHCO₃) | 1,000 ppm HCO₃⁻; 377 ppm Na⁺ | Vacuum brewer, triangle tests | Lockhart et al.7 |
| Carbonate threshold in water and in coffee (as Na₂CO₃) | 44 ppm in water; 125 ppm in coffee | As above | Lockhart et al.7 |
| Highest bicarbonate in the US supplies surveyed (1955) | 336 ppm | Below the threshold in water | Lockhart et al.7 |
| Iron, visible colour change | Detectable at 1 ppm side by side; clearly greenish above 7 ppm | Coffee with cream | Lockhart et al.7 |
| Bicarbonate range tested, 2026 | Up to 300 mg/L | Colombian coffee, duo-trio and descriptive panels | Rune et al.10 |
| Waters of a 2021 filter study, total hardness | 9.33, 37.33, 71.67, 157.50 mg/L as CaCO₃ | 6 g coffee per 100 mL, filter machine | Dadalı & Elmacı9 |
| Same waters, dissolved solids | 24.00, 59.67, 141.33, 424.33 mg/L | As above | Dadalı & Elmacı9 |
| Sourness, softest to hardest water | 3.35, 2.88, 2.55, 1.52 (0–5 scale) | Panel of 10 | Dadalı & Elmacı9 |
| Bitterness, softest to hardest water | 1.30, 1.55, 2.82, 3.63 (0–5 scale) | Panel of 10 | Dadalı & Elmacı9 |
The four waters of the 2021 study were purified water, bottled water, and medium-hard and hard mains water.9 Its abstract gives the highest sourness as 3.55; its table gives 3.35 for the purified water. We use the table. The 1955 thresholds come from tests with sodium salts. For sodium bicarbonate the authors could not assign the effect to either ion; for sodium carbonate they attributed it almost entirely to carbonate.7 Both handbooks place the flow effect at high carbonate or alkalinity levels; they state their thresholds in different quantities.1,2
Acid neutralisation#
According to a coffee science textbook chapter, the most commonly cited finding is that perceived acidity correlates with titratable acidity; the acids in coffee entry covers the exceptions.3 The chapter argues from this that the acid the tongue senses is the acid extracted from the coffee minus what the water’s alkalinity has neutralised.3 On its own measurements, water of moderate alkalinity accounted for a substantial share of the titratable acidity of filter coffee and a much smaller share in lungo espresso.3 We do not give its percentages because no second source carries them.
By our calculation, alkalinity of 50 ppm as CaCO₃ equals 1 mmol/L of acid-neutralising capacity, because one CaCO₃ (100.1 g/mol) neutralises two protons.1,3 A litre of such water can therefore neutralise 1 mmol of a monoprotic acid. At 75 ppm, the upper end of the SCA range, it is 1.5 mmol.
The reaction releases carbon dioxide. At high alkalinity, the SCA handbook and the textbook chapter describe this gas as an added resistance in the coffee bed.1,3 According to the chapter, high alkalinity with low hardness also raises pH on heating, more so with sodium.3 A suggested consequence is lower solubility of carbohydrates and more resistance during extraction.3 The same chapter reports that high hardness with equal or lower alkalinity did not lengthen percolation time.3
Cations and extraction#
The 2014 model used density functional theory to calculate binding energies of sodium, magnesium and calcium ions to five coffee acids, caffeine and the aroma compound eugenol.5 Its authors use these energies to propose a mechanism and an ideal mineral composition.5 A calculated binding energy is not a measurement of extraction from a coffee bed. This entry treats its conclusions as hypotheses.
Brewing trials give a mixed picture:
- Refractometer. In the SCA’s laboratory tests, harder water did not raise extraction measurably.1
- Acids. In a 2024 study, magnesium and calcium salts at drinking-water levels changed citric, malic, lactic and quinic acids little; only ten-fold higher levels gave clearer differences.8 Adding the salts before or after brewing gave similar acid contents in most cases.8
- Commercial waters. In a 2024 cold-brew study, calcium, magnesium and bicarbonate correlated positively with extraction.11 The six waters differed in several ions at once,11 so the correlation does not isolate a cause.
- Single salts. An older study summarised in the SCAA handbook found a stronger brew with sodium bicarbonate than with calcium chloride.2 This is the opposite of what a simple “hardness extracts more” rule would predict.
The 2024 acid study also notes that interactions after brewing may still influence perceived flavour.8 Its authors call a scientific basis for water recommendations long overdue.8
Sensory studies compared#
| Study | Brew | Waters | Main finding | Source |
|---|---|---|---|---|
| Lockhart et al., 1955 | Vacuum brewer | Single salts at taste threshold and above | No difference at threshold; higher levels needed in coffee | 7 |
| Cupping trials in a textbook chapter | Cupping | Three waters, low to high in hardness and alkalinity | Lowest-mineral water scored highest for most attributes in two coffees | 3 |
| Eo & Kim, 2017 | Not stated in abstract | Five commercial waters and distilled water | Significant differences in flavour, acidity, bitterness, sweetness, body and aftertaste; bitterness correlated with magnesium | 12 |
| Dadalı & Elmacı, 2021 | Filter machine | Four waters, 9–158 mg/L hardness | Sourness down and bitterness up with harder water | 9 |
| Chen et al., 2024 | Cold brew | Six commercial waters | Lower ion levels favoured fruity and floral notes; moderate levels rated best overall | 11 |
| Rune et al., 2026 | Brewed coffee | Calcium, magnesium and bicarbonate varied separately | Alkalinity, not hardness, drove the sensory changes | 10 |
Two patterns run through the table. Where alkalinity was varied, perceived acidity fell as it rose.1,10 Where waters were ranked by total mineral content, results differ. In the cupping trials the lowest-mineral water scored highest,3 and in the 2017 study rated quality fell in a straight line as total mineral content rose.12 In the cold-brew study, lower ion levels favoured fruity and floral notes, but moderate ion levels gave the best overall rating.11
The textbook chapter cautions that its low-mineral result may reflect the high extraction typical of cupping; another brew method might rank the waters differently.3 It also found that the two coffees responded differently to the same waters.3
Open questions#
- Calcium and magnesium separately. The textbook chapter recommends measuring calcium and magnesium individually instead of total hardness if their sensory effects prove different.3
- Coffee by water interactions. The same water change did not affect every coffee in the same way.3 No study opened for this entry tested many origins and roast degrees with the same set of waters.
- Roast style. The SCA reports only anecdotal evidence that harder, more alkaline water can partly compensate for underdeveloped roasts.1
- Aroma. According to the textbook chapter, the effect of water pH and specific substances on aroma binding and volatility is unexplored.3
- Mechanism of the 2026 calcium effect. Small differences in samples with both calcium and bicarbonate were attributed to calcium carbonate precipitating and lowering the effective bicarbonate, not to calcium itself.10
Key studies#
Rune et al., 2026 – Water composition and coffee sensory properties. Varied hardness (calcium, magnesium) and alkalinity (bicarbonate) in brewing water and tested the coffees with duo-trio and descriptive sensory analysis. Hardness alone produced no marked flavour change; bicarbonate up to 300 mg/L lowered acidity, fruitiness and complexity and raised roasted and bitter notes.
Hendon et al., 2014 – Cations and extraction. Calculated, with density functional theory, how strongly sodium, magnesium and calcium bind five coffee acids, caffeine and eugenol. The results are used to argue for a mechanism and an ideal mineral composition; they are a model, not a measurement of extraction.
Bratthäll et al., 2024 – Divalent cations and coffee acids. Added magnesium and calcium salts before or after brewing and measured organic acids by GC–MS and NMR. Drinking-water levels changed the acids little; ten-fold higher levels changed them more.
Dadalı & Elmacı, 2021 – Hardness and filter coffee. Brewed filter coffee with four waters from 9 to 158 mg/L hardness and measured volatiles and sensory intensities. Sourness was highest with the softest water and bitterness highest with the hardest; alkalinity was not measured.
Lockhart et al., 1955 – Water impurities and flavour. Measured taste thresholds of common ions in water, then brewed coffee with water at and above those levels. At threshold, panels could not tell the coffee from coffee made with distilled water; iron mainly changed the colour.
- Hardness and extraction yield: a density-functional model of how cations bind coffee compounds5 is read by a textbook chapter as mounting evidence that calcium and magnesium drive extraction;3 the SCA handbook by the same authors reports laboratory tests in which a refractometer showed no effect.1 A 2024 study found little change in four acids at drinking-water salt levels,8 while a 2024 cold-brew study across six commercial waters found calcium, magnesium and bicarbonate positively correlated with extraction.11 An older study reported in the SCAA handbook gave a weaker brew with calcium chloride than with sodium bicarbonate.2
- Magnesium versus calcium: the SCA handbook cites theory and preliminary experiments for slightly higher extraction with magnesium but calls the sensory difference unclear;1 a 2017 correlation study found bitterness highly correlated with magnesium content.12 No study opened for this entry compared the two at equal hardness in a brewing trial with a refractometer.
- How much bicarbonate the cup can detect: in 1955 triangle tests, sodium bicarbonate was first detectable in coffee at about 1,000 ppm bicarbonate;7 a 2026 study found clear sensory changes as bicarbonate rose up to 300 mg/L.10 The two studies used different test designs: triangle tests in 1955, duo-trio and descriptive analysis in 2026.7,10
- Which water scored best: in cupping tests reported in a textbook chapter, the water lowest in hardness and alkalinity scored highest for most attributes;3 a 2021 filter-coffee study recommended medium-hard water because it gave the most volatile compounds.9
Coffearium (2026). Water minerals and flavour. Coffearium – The open coffee compendium. Retrieved 22 October 2026, from https://coffearium.org/entries/water-minerals-and-flavour/
@misc{coffearium2026-water-minerals-and-flavour,
author = {{Coffearium}},
title = {{Water minerals and flavour}},
year = {2026},
howpublished = {Coffearium -- The open coffee compendium},
url = {https://coffearium.org/entries/water-minerals-and-flavour/},
urldate = {2026-10-22},
note = {Retrieved 22 October 2026}
}Entries are updated when the evidence changes; the retrieval date tells your readers which version you used. Changes are listed in the Change log.
Change log
- – Removed the flow thresholds "> 100 ppm as CaCO₃" (SCA handbook) and "> 100 mg/L" (SCAA handbook) from the quick facts and reference table. Reason: Each rests on one Tier B handbook passage (the SCA figure is repeated only by a textbook chapter by the same authors); the number rule needs one Tier A or two independent Tier B sources.
- – Corrected the 1955 note; the authors attributed the sodium carbonate effect almost entirely to carbonate, and could not separate the ions only for sodium bicarbonate. Reason: Lockhart et al. 1955 (full text).
- – Chen et al. 2024 waters described as commercial, not bottled; Eo & Kim 2017 described as "highly correlated" rather than "most strongly"; removed the claim that the 2021 study's recommendation did not reflect panel preference. Reason: Wording went beyond the abstracts.
Sources
- A/BWellinger, M.; Smrke, S.; Yeretzian, C. (2018). The SCA Water Quality Handbook, edition 2 (ISBN 978-0-9995807-3-8). Specialty Coffee Association.
- A/BBeeman, D.; Songer, P.; Lingle, T. (ed. Everage, L.) (2011). The Water Quality Handbook (ISBN 978-1-882552-08-5). Specialty Coffee Association of America.
- BWellinger, M.; Smrke, S.; Yeretzian, C. (2017). Water for Extraction – Composition, Recommendations, and Treatment. In: Folmer, B. (ed.), The Craft and Science of Coffee, Academic Press, ch. 16, pp. 381–398. doi:10.1016/B978-0-12-803520-7.00016-5 · other copies
- BMestdagh, F.; Glabasnia, A.; Giuliano, P. (2017). The Brew – Extracting for Excellence. In: Folmer, B. (ed.), The Craft and Science of Coffee, Academic Press, ch. 15, pp. 355–380. doi:10.1016/B978-0-12-803520-7.00015-3 · other copies
- AHendon, C. H.; Colonna-Dashwood, L.; Colonna-Dashwood, M. (2014). The Role of Dissolved Cations in Coffee Extraction. Journal of Agricultural and Food Chemistry 62(21), 4947–4950. doi:10.1021/jf501687c · other copies
- ANavarini, L.; Rivetti, D. (2010). Water quality for Espresso coffee. Food Chemistry 122(2), 424–428. doi:10.1016/j.foodchem.2009.04.019 · other copies
- ALockhart, E. E.; Tucker, C. L.; Merritt, M. C. (1955). The Effect of Water Impurities on the Flavor of Brewed Coffee. Journal of Food Science 20(6), 598–605.
- ABratthäll, T.; Figueira, J.; Nording, M. L. (2024). Influence of divalent cations on the extraction of organic acids in coffee determined by GC-MS and NMR. Heliyon 10(5), e26625. doi:10.1016/j.heliyon.2024.e26625 · other copies
- ADadalı, C.; Elmacı, Y. (2021). The Effect of Water Hardness on Volatile Compounds and Flavour of Filter Coffee. Turkish Journal of Agriculture – Food Science and Technology 9(2), 312–320. doi:10.24925/turjaf.v9i2.312-320.3885 · other copies
- ARune, C. J. B.; Steen, I.; Kjærgaard, E. R.; Maroun, Z.; Bredie, W. L. P.; Kjaergaard, H. G. (2026). Impact of Water Composition on Coffee Sensory Properties: A Comprehensive Analysis. Journal of Food Science 91(9), e71434. doi:10.1111/1750-3841.71434 · other copies
- AChen, S.; Xiao, Y.; Jiang, F.; Jiang, T.; He, X.; Zhu, J.; Tang, W.; Zhou, Y. (2024). Effects of Water Quality Factors on the Main Components and Flavor of Cold Brew Coffee. Science and Technology of Food Industry (Shipin gongye ke-ji) 45(5), 70–80. doi:10.13386/j.issn1002-0306.2023050076 · other copies
- AEo, H.-J.; Kim, J.-S. (2017). Analysis of Correlations between Mineral Contents in Waters and Sensory Characteristics of Coffee. Culinary Science and Hospitality Research 23(4), 105–115. doi:10.20878/cshr.2017.23.4.011 · other copies
Tier A · authoritative · B · trade and expert · C · general. Supporting quotes for every source are in the data version of this entry. If a publisher’s site refuses access, try the doi link or “other copies”.
Discussion
A space for readers to share experience from the farm, the roastery or the bar, ask each other and offer other views. The Coffearium team does not reply here. Please keep to the community rules; to change the entry itself, use Suggest a change.