Water for brewing coffee

A cup of brewed coffee is roughly 98–99% water by weight.¹ Everything you taste is dissolved in something you probably never chose. Two people can buy the same bag, weigh the same dose, use the same brewer and the same ratio, and get two different drinks — because their taps are different.

Water is not a neutral solvent. The minerals already dissolved in it do two jobs at once: some of them help pull flavour out of the grounds, and some of them cancel out the acidity of what comes out. Get the balance wrong and no amount of grinder tuning fixes it.

The five numbers

Water for coffee gets described with the same short list of measurements everywhere. All of them except pH are quoted in parts per million (ppm), which for water is the same as milligrams per litre (mg/L).

What it’s called	What it actually means	Why coffee cares
TDS (total dissolved solids)	how much mineral is in the water in total	the headline number; a $15 meter reads it
Hardness	how much calcium and magnesium specifically	these are the ions that help dissolve flavour compounds out of the grounds1
Alkalinity	how much bicarbonate — the water’s ability to neutralise acid	this is what mutes or brightens the cup2
pH	how acidic or alkaline the water is right now	least useful of the five — see below
Chlorine / chloramine	disinfectant your utility added	must be zero

Two of these get confused constantly. Hardness and alkalinity are not the same thing, and they are not measured on the same scale even though both are conventionally written “as CaCO₃” — as if the mineral were all chalk. Hardness is the calcium-and-magnesium half; alkalinity is the bicarbonate half. Water can be hard and low in alkalinity, or soft and high in it.

pH matters less than people assume. Alkalinity is buffering capacity — how much acid the water can absorb before its own pH moves — and it is capacity, not starting pH, that decides what happens when hot water meets a few grams of acidic coffee.²

The word TDS is also doing double duty on this wiki. Here it means minerals in the water before brewing; on the Total dissolved solids note it also means dissolved coffee in the finished cup. Same three letters, opposite ends of the same brew.

Alkalinity is the lever that moves flavour

The clearest single claim in this area comes from the roasting consultant Scott Rao: “Alkalinity is the single-most important factor in how water will affect coffee flavor.”² The mechanism is simple arithmetic — bicarbonate neutralises acid, so:

• Lower alkalinity → coffee tastes more acidic, brighter, more like fruit.
• Higher alkalinity → coffee tastes less acidic, flatter, duller.

Rao states these hold true independently of the water’s pH.² His own preference for lightly roasted, well-developed coffee is an alkalinity of 30–40 ppm, while allowing that this is taste, not law.²

This is why the same coffee tastes thrilling in one city and boring in another. It is not always the barista.

On calcium versus magnesium — which of the two hardness minerals is better, and in what proportion — Rao is blunt that “there is little consensus on the optimal level or the ideal balance of Ca+ and Mg+.”² Practitioners lean magnesium-forward for light roasts, but that is preference, not settled science.

The one peer-reviewed paper, and the book behind it

Almost every water guide in specialty coffee eventually points at the same two documents, both from the same pair of authors.

The paper is “The Role of Dissolved Cations in Coffee Extraction,” by Christopher H. Hendon, Lesley Colonna-Dashwood and Maxwell Colonna-Dashwood, published in the Journal of Agricultural and Food Chemistry in 2014 (volume 62, pages 4947–4950).³ Hendon is a computational chemist; Maxwell Colonna-Dashwood is a Bath café owner and repeat UK Barista Champion who has placed as high as fifth at the World Barista Championship. That combination — a lab and a competition bar — is why the work landed the way it did.

The book is Water for Coffee, by Maxwell Colonna-Dashwood and Christopher H. Hendon, first published in 2015.⁴ It is the closest thing the industry has to a reference text on the subject, and it is the ultimate source of most of the hardness-and-alkalinity explanations circulating on blogs, usually several hands removed.

Worth being clear about what this changes and what it doesn’t. It gives the field one genuinely peer-reviewed anchor, which is more than most brewing arguments have. It does not settle the calcium-versus-magnesium question in practice — see the callout below.

Chlorine has to go

Municipal water is disinfected with chlorine or chloramine. Both wreck coffee: they react with coffee’s phenolic compounds to produce chlorophenol off-flavours — the medicinal, swimming-pool note.⁵ The Specialty Coffee Association’s brewing water description reportedly requires total chlorine of zero.⁶

The fix is carbon filtration, which strips chlorine and chloramines without removing the minerals you want to keep.⁷ Letting water stand in an open jug also works — but only for chlorine, which gasses off; chloramine is deliberately more stable and does not.

What the SCA actually says — and doesn’t

Every guide to brewing water quotes “the SCA standard.” The numbers most often quoted are a TDS target around 150 mg/L, an alkalinity around 40 mg/L as CaCO₃, a pH of 6.5–7.5, sodium below 30 ppm, and no detectable chlorine.⁷¹⁸⁶

Those centre values are consistent across every source consulted here. The ranges around them are not, and the SCA’s own public standards page publishes no water values at all — see the callouts below before quoting any of it.

One specific misattribution is worth killing. Guides and videos regularly cite a document called “SCA Standard 310-2021, Water Quality.” The SCA’s own standards listing was fetched for this note, and SCA-310 is “Home Coffee Brewers: Specifications and Test Methods” — a standard about brewing machines, not about water.⁹ The full published list runs SCA-102 to SCA-105 and SCA-710 for value assessment and evaluator competencies, and SCA-310, 320, 350 and 510 for home brewers, home grinders, espresso machines and training venues. Nothing on it concerns water. So when a source hands you “310-2021” as the citation for a TDS number, it has attached a real standard number to the wrong subject, and whatever chain it copied that number from is not a chain anyone should trust.

Making your own water

The mainstream approach in specialty coffee is to start from nothing and add back: strip the water to near-zero minerals with reverse osmosis (RO) or buy distilled, then dose in exactly the hardness and alkalinity you want. This removes local variation entirely, which is why cafés and roasteries do it.

Two documented recipes, both from single authors rather than standards bodies:

A magnesium-forward brew water. Make two stock concentrates, then dilute:⁸

• Concentrate A (hardness): 2.45 g magnesium sulfate — Epsom salt — in 500 mL distilled water.
• Concentrate B (buffer): 1.68 g sodium bicarbonate — baking soda — in 500 mL distilled water.
• To brew: add 50 mL of A and 17 mL of B to one US gallon (3,785 mL) of distilled water, giving roughly 50 mg/L magnesium hardness and 20 mg/L alkalinity. For darker roasts, raise B to 25 mL.

A calcium-forward brew water. A four-salt concentrate meant to imitate a natural mineral profile rather than a magnesium-only one. Dissolve in 500 mL of pure water: 0.37 g calcium chloride dihydrate, 0.26 g calcium sulfate dihydrate, 0.17 g sodium bicarbonate and 0.14 g sea salt. At full strength — 25 mL of concentrate into 475 mL of pure water — that lands at 25 ppm hardness from the chloride, 15 ppm from the sulfate, and 10 ppm alkalinity from the baking soda. Half strength is 12.5 mL into 487.5 mL.¹⁰

Notice the shape of it: about 40 ppm total hardness against only 10 ppm alkalinity. Very low buffer, deliberately — that is a recipe built to let acidity through.

“Rpavlis” water, for espresso. Dissolve 0.1 g potassium bicarbonate in 1 litre of distilled or RO water. That gives zero general hardness and about 50 ppm alkalinity — so it cannot form scale at all, while still having enough buffer to tame a sharp shot.¹⁰ The logic is worth stating plainly: scale needs calcium and magnesium, and this water contains neither, so there is nothing available to precipitate onto a boiler wall no matter how long it runs.

For anyone not weighing salts, the same source suggests the practical middle path: RO water blended back with a little tap water, or RO plus a mineral concentrate.¹⁰ Its working targets are 30–100 ppm TDS for filter and 70–120 ppm for espresso — soft, on the grounds that low-mineral water is the safer error.¹⁰

Scale, corrosion, and the compromise nobody escapes

Water that is best for flavour is not automatically best for the machine. Hardness plus heat produces scale — the mineral crust that kills boilers and blocks group heads. Push hardness too low, though, and water becomes aggressive toward metal instead.

Rao’s framing of the trade-off is the honest one: “Sometimes one needs to compromise and allow a little scale or sacrifice a little flavor quality to balance those two concerns.”² This is why café water systems are usually RO with a bypass — a deliberate trickle of untreated water blended back in — rather than pure RO.

Test, don’t guess

The single most repeated piece of advice, and the cheapest: find out what is actually in your water before you treat it.²⁷⁸ In practice that means a municipal water-quality report for your city, aquarium test strips for general hardness (GH), carbonate hardness (KH) and pH, a cheap TDS meter for the headline number, or a proper lab analysis if you are specifying equipment for a business. Rao warns specifically against relying on free testing offered by the vendor selling you the filter.²

A filter chosen without data is a guess with a subscription.

The other water bill

Everything above concerns the half-litre that meets the grounds. That is the smaller number by an enormous margin.

Growing the beans for one cup of coffee takes about 140 litres of water on average, and producing 1 kg of roasted coffee in Africa, South America or Asia takes roughly 26,400 litres.¹¹ That is the agricultural water footprint — rainfall the trees drink plus irrigation and processing — not anything a brewer controls.

The two numbers belong on the same page because they are the same resource seen at two scales. A café can obsess over 40 ppm of magnesium in a jug while the coffee in that jug already cost a bathtub of water at origin. It is also why washed processing, which uses water to strip the fruit off the seed, is a live sustainability question in dry-season origins, and why shifting rainfall is the threat producers name first.

Where this shows up on the wiki

Competition water is far softer than any of the guidance above:

• Nas Jaafar, the 2026 World Brewers Cup champion, competed on water at 50 ppm.
• Tetsu Kasuya used water at 0.33 ppm in 2016 — effectively distilled.

Both sit at or below the bottom of every published recommendation. That gap is the point: the published numbers describe water that works across a whole menu on a busy bar, day after day, without destroying the machines. A competitor is optimising one coffee, once, on equipment that only has to survive fifteen minutes.

The same logic runs through Cupping, where standardised water exists so that scores from different rooms can be compared at all, and through Pour-over brewing, where the water is the only ingredient a home brewer never thinks to change.

To verify

The SCA’s water numbers, and whether there is a current public standard at all. Two incompatible ranges circulate. One set gives TDS 50–175 ppm with total hardness 17–85 ppm;⁷ another gives TDS 75–250 mg/L with hardness 50–175 mg/L.¹⁸ Notice that 50–175 is TDS in the first and hardness in the second — one of these is very likely a transcription error propagating between guides, and this wiki cannot say which. Worse, the Specialty Coffee Association’s public standards page was fetched for this note and mentions water nowhere at all — it lists standards for value assessment, equipment, green grading and roast levels only.⁹ The only source that names a document at all calls it “SCA Standard 310-2021, Water Quality” — and that number belongs to the home-brewer standard instead, as the SCA’s own listing shows.⁹ Every other attribution found was a bare link to sca.coffee/research with no document title.⁷ Treat every “SCA water standard” figure on this page as a secondary reading of a document nobody in this chain opened, and quite possibly of a document that does not exist in the form quoted.

To verify

What the 2014 cations paper actually concluded. The bibliographic facts are solid — title, three authors, journal, volume 62, pages 4947–4950, year 2014 — all confirmed against the publisher’s own registered metadata.³ The findings are not. The research behind this note claims the paper shows magnesium enhances extraction more than calcium at equal hardness, and that bicarbonate suppresses acidity. That claim is repeated everywhere, but the publisher has paywalled the article and even elided the abstract from the open citation databases, so no wording from the paper itself was read for this note.³ Treat the magnesium-beats-calcium summary as the coffee industry’s folk memory of a paper, not as a quotation from it — and note that Scott Rao says flatly there is still “little consensus” on the Ca/Mg balance,² which is not what you would expect if one paper had settled it a decade ago.

To verify

The 99.4% / 97.8% water figures. The research behind this note attributes to Wikipedia’s coffee article the claims that brewed coffee is 99.4% water and restaurant espresso 97.8% water. That article was fetched here and states neither figure.¹¹ The broader “98–99% water” range on this page comes from a different source and stands.¹ The two precise percentages should not be quoted until someone finds where they actually came from.

To verify

Whether very soft water over-extracts or under-extracts. Sources contradict each other head-on. One states that too-soft water under-extracts, giving “a thin, sour, flat cup,” for lack of the ionic charge needed to dissolve flavour compounds.⁷ The report behind this note cites another guide claiming water below 50 ppm over-extracts and tastes harsh¹² — the same claim already flagged on the Total dissolved solids note.¹³ These are opposite predictions from the same input. The mechanism described in the first is at least chemically coherent, but neither is traced here to a controlled study, and both compress hardness, alkalinity and TDS into a single number that cannot carry all three.

To verify

The sensory ceiling. The claim that water above roughly 200 ppm produces muted, chalky cups is widely repeated and appears in the research behind this note,¹² and one fetched source agrees directionally that too-hard water gives “a harsh, chalky, or dull brew.”⁷ It is an expert generalisation repeated across guides, not a published finding from the SCA, Coffee Quality Institute or any study read here.

To verify

The recipe outputs are calculated, not measured. The “50 mg/L hardness, 20 mg/L alkalinity” result of the magnesium concentrate recipe, and the “0 GH / 50 ppm KH” result of the Rpavlis recipe, are both the authors’ own arithmetic from the salt weights.⁸¹⁰ The chemistry is plausible and the recipes are widely used, but no lab analysis confirming either output was located. Also note the first recipe lands at 20 mg/L alkalinity — only half the commonly quoted 40 mg/L target — and its author describes that as working deliberately well for light roasts.⁸ Low-alkalinity brewing water is a stylistic choice, not an error.

Related

• Total dissolved solids — the same three letters, measured on the finished cup instead of the water
• Extraction yield — what the minerals in the water are actually changing
• Maxwell Colonna-Dashwood · Scott Rao — the two names most of this page’s claims trace back to
• Washed process · Climate change and coffee — water as an agricultural input rather than a brewing one
• Brew ratio · Coffee grinding — the other two variables people tune first, usually before water
• Pour-over brewing — where soft, deliberate water shows up most
• Cupping — evaluation depends on standardised water existing
• Specialty Coffee Association — the body every water number gets attributed to
• Coffee Quality Institute — the other standards body in the room
• World Brewers Cup · Nas Jaafar · Tetsu Kasuya — competition water on the record
• Coffee roasting — roast level changes how much alkalinity you want
• Third wave coffee — the movement that made water a variable worth arguing about

Official links

• Official site — Specialty Coffee Association, Coffee Standards

Footnotes

1. All About Coffee — Coffee Water Chemistry — confirmed via WebFetch; source of the 98–99% figure and the chlorophenol mechanism ↩ ↩² ↩³ ↩⁴ ↩⁵

2. Scott Rao — Demystifying Water for Coffee — confirmed via WebFetch; source of all quoted sentences ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰

3. Hendon, Colonna-Dashwood & Colonna-Dashwood — “The Role of Dissolved Cations in Coffee Extraction,” Journal of Agricultural and Food Chemistry 62, 4947–4950 (2014) — bibliographic record confirmed against the publisher’s registered Crossref metadata; the article itself is paywalled and its abstract is elided from open databases ↩ ↩² ↩³

4. Water for Coffee — Maxwell Colonna-Dashwood and Christopher H. Hendon (2015, Independent Publishing Network) — publication details confirmed via Open Library ↩

5. All About Coffee — Water Chemistry Basics — cited in the research report for the hardness definitions ↩

6. How To: Brew Water Chemistry for Coffee & Espresso — cited in the research report as presenting SCA’s water standard, not independently fetched ↩ ↩²

7. Coffeester — Water for Coffee: TDS, Hardness, Alkalinity & the SCA Water Standard — confirmed via WebFetch ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷

8. Brewed Barista — Water Chemistry for Coffee Brewing: The Complete Guide — confirmed via WebFetch; source of the concentrate recipe ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶

9. Specialty Coffee Association — Coffee Standards — fetched for this note; contains no water values, and lists SCA-310 as “Home Coffee Brewers: Specifications and Test Methods” ↩ ↩² ↩³

10. The Coffee Chronicler — Everything You Should Know About Water for Coffee — confirmed via WebFetch; source of the Rpavlis recipe and the filter/espresso TDS targets ↩ ↩² ↩³ ↩⁴ ↩⁵

11. Coffee — Wikipedia — fetched for this note; source of the 140 L per cup and 26,400 L per kg water-footprint figures ↩ ↩²

12. My Coffee Explorer — Water Chemistry for Coffee: The Invisible Ingredient — cited in the research report; returned HTTP 404 when fetched, so its wording could not be checked ↩ ↩²

13. Itadi Coffee — The Ultimate Guide to Coffee Water: pH, TDS, and Mineral Content — the source of the same over-extraction claim already flagged on the Total dissolved solids note; not independently fetched ↩