A brewed coffee is roughly 98.5% water, and espresso around 90%, which makes water by far the largest ingredient. Baristas have treated it as a variable for years, and specialty cafés routinely install treatment systems or mix their own brewing water from concentrates. Whether that is worth doing at home is a fair question, and it now has a better answer than it did, because two pieces of research approach it from opposite directions: one from theoretical chemistry, one from a sensory panel. They agree on more than they disagree, and where they diverge is the practically useful part.
What is actually dissolved in water
Tap water contains dissolved minerals, and two measures describe the part that matters for coffee. Hardness is the concentration of divalent cations, mainly calcium and magnesium. Alkalinity is the water's capacity to neutralise acid, which in drinking water comes mostly from bicarbonate.
The two often rise together, because the same limestone geology contributes both, which is why they are easy to confuse and why advice about hard water often conflates them. They act on coffee through different mechanisms, and the research below separates them.
The chemistry: why minerals should matter
Hendon and colleagues[1] approached this computationally. The flavour compounds in coffee include charge-neutral species as well as acids and their conjugate salts, and their dissolution and extraction depend on the dissolved mineral content of the water, since cations such as sodium, magnesium and calcium coordinate to nucleophilic sites on coffee molecules. Using density functional theory, they quantified binding energies for five coffee acids, caffeine and eugenol, a representative flavour compound, and from that derived insight into the mechanism and an ideal mineral composition for extracting flavoursome compounds.
The practical message that circulated from that work is that magnesium binds these compounds more effectively than calcium, so magnesium-rich water should extract more of the flavour compounds people want. That is a calculation about binding rather than a brewing experiment, and it should be read as a mechanism rather than a recommendation confirmed in cups.
The sensory test: what people could actually taste
Rune and colleagues[2] ran the experiment the chemistry implies. Using a Colombian Kachalu coffee, they compared brews made with ultrapure water as a reference against brews from ultrapure water dosed with calcium, magnesium and bicarbonate, spanning from the lower end of the commonly cited ideal brewing zone up to deliberately extreme levels. Evaluation used duo-trio screening, where panellists identify which of two samples matches a reference, and descriptive sensory analysis.
The headline result reorders the usual advice. Across all sensory evaluations, water hardness alone produced limited effects, and alkalinity, not hardness, was the primary determinant of sensory change. Even small increases in bicarbonate made the coffee less acidic and fruity and more flat or bitter. Variations in calcium within realistic ranges showed little sensory relevance.
Their practical conclusion was that controlling alkalinity may be more important than adjusting hardness when designing water for brewing, and that cafés, consumers and equipment manufacturers can use this to improve consistency.
Why bicarbonate does that
The mechanism is straightforward once stated. Coffee's perceived brightness comes substantially from its organic acids, and bicarbonate neutralises acid. Water with high alkalinity buffers the brew, flattening the acidity that a lighter roast is prized for and shifting the balance toward bitterness.
That fits what the acid research shows about where coffee's taste comes from. Rune and colleagues[3] measured acids in brewed coffee against sensory thresholds and found perceived acidity is not the work of any single acid, and Santanatoglia and colleagues[4] found chlorogenic acids had the largest influence on pH and titratable acidity while organic acids tracked perceived sourness. Buffering that system is a blunt intervention, and the sensory panel detected it.
The ideal brewing zone, and where it came from
Specialty coffee guidance has long recommended a target range for brewing water, typically framed as a total dissolved solids figure with separate limits for hardness and alkalinity. That framework predates the studies here, and the experiments in Rune and colleagues[2] were designed partly to test its edges, spanning the lower end of the commonly cited ideal zone up to deliberately extreme levels.
What their results suggest is not that the guidance is wrong but that it is weighted oddly. A recommendation that gives hardness and alkalinity similar prominence does not reflect a finding that one of them drove nearly all the detectable sensory change. For anyone choosing water rather than formulating it, that simplifies the task considerably: one number to check instead of two.
Magnesium versus calcium
The distinction between the two hardness minerals is a staple of barista discussion, and it comes almost entirely from Hendon and colleagues[1], whose calculated binding energies favoured magnesium. It is a reasonable prediction and it has driven a small industry of magnesium-based brewing water concentrates.
The sensory evidence is more restrained. In Rune and colleagues[2], hardness alone produced limited effects across the evaluations, and calcium variation within realistic ranges had little sensory relevance. That does not refute the chemistry, which describes extraction mechanisms rather than predicting panel scores, and the two studies were asking different questions. It does suggest that a home brewer choosing between two waters should look at the bicarbonate figure before comparing magnesium content.
Where water sits among the other variables
It is worth keeping the size of this effect in perspective. Extraction is governed by grind size, temperature, time, ratio and agitation, and those levers are large. Cameron and colleagues[5] modelled espresso systematically and showed how much extraction yield and reproducibility depend on grind and dose, to the point of recommending counterintuitive changes such as using less coffee.
Water is a real variable that sits alongside those, not above them. If your coffee tastes bad and your grinder is a blade grinder, the water is not the problem, and our grinder guide covers why.
There is also a practical floor, which is that water with off-flavours makes bad coffee regardless of mineral content. Chlorine is the usual culprit in treated tap water, and a basic carbon filter removes it, which is the cheapest meaningful upgrade available.
What about distilled or ultrapure water?
The temptation, given that minerals can spoil things, is to remove them all. That is a mistake on both the chemistry and the sensory evidence. Hendon and colleagues[1] describe cations as participants in extraction rather than contaminants, so water with nothing dissolved in it extracts differently and, by most accounts, worse. Ultrapure water served as a reference point in Rune and colleagues[2] rather than as a recommendation.
Distilled and reverse-osmosis water are also aggressive toward metal equipment and are explicitly discouraged for espresso machines, which is a maintenance issue rather than a flavour one but a real cost.
Reading this from the Philippines
Tap water quality varies considerably between areas here, and bottled purified and mineral water are widely used for drinking and for home brewing. That makes the alkalinity finding unusually actionable, because bottled water labels normally print mineral composition, and the bicarbonate figure is the one this research says to look at. A purified water with low mineral content sits at one end, a mineral water with high bicarbonate at the other, and the sensory study suggests the second will flatten a bright coffee.
For a café, the same logic applies with equipment attached, since alkalinity also drives scale formation in boilers. For someone brewing at home, the low-cost experiment is to make the same coffee with two different bottled waters and see whether the difference is audible to you. Given the panel in Rune and colleagues[2] detected changes from small bicarbonate increases, there is a reasonable chance it will be.
Temperature, scale and the machine
Water does one more thing that has nothing to do with flavour chemistry: it deposits scale. Calcium and bicarbonate together produce the limescale that clogs boilers, thermoblocks and group heads, and the water that is worst for that is also the water the sensory study identified as worst for taste. That is a rare alignment in equipment advice, where flavour and maintenance usually pull in opposite directions.
It also explains why cafés treat water at all, since a descaling schedule and a failed boiler are more expensive than a filter cartridge. For home brewers with an espresso machine, the same logic applies at a smaller scale, and it is often the reason people start paying attention to water in the first place, with the flavour benefit arriving as a side effect.
What a home experiment should control
If you want to test this yourself, the design matters, because water is a subtle variable next to the loud ones. Hold the coffee, grind setting, dose, ratio, temperature and brew time constant, and change only the water, ideally brewing both within a few minutes of each other so the coffee is equally fresh. Tasting them side by side rather than on consecutive days is what makes a small difference detectable, and it is what the duo-trio method in Rune and colleagues[2] formalises.
The coffee you choose matters too. A bright, lightly roasted single origin has more acidity for alkaline water to flatten, so any difference will be easier to notice than with a dark roast where those notes are already gone.
Why cafés bother and most homes need not
A café is solving a different problem from a home brewer. It brews hundreds of cups a day on equipment that must behave identically each time, its water arrives from a single supply whose composition shifts with season and source, and its espresso machine is an expensive appliance that scale will eventually destroy. Controlling water buys consistency and protects hardware, and both are worth paying for at that scale.
At home, the same reasoning applies weakly. A filter jug and a sensible bottled water cover most of the benefit, and the difference between a carefully formulated brewing water and a reasonable one is smaller than the difference a fresher bag of coffee makes. Water is worth attending to once grind, ratio and freshness are settled, which is the order the evidence supports rather than the order the internet usually presents.
The limits of this evidence
The computational study calculated binding energies rather than brewing anything, and its conclusions about ideal water are inferences from those calculations. The sensory study used one coffee, a single Colombian lot, with one brewing method, and a coffee's response to water may depend on its own acid profile, so a dark roast might behave differently. Trained sensory panels detect differences that ordinary drinkers may not, and neither study tested consumer preference at scale. And none of this work addressed espresso specifically, where the pressures and concentrations differ.
It is also worth noting how new the sensory side of this is. The chemistry argument has circulated in specialty coffee since 2014, and the controlled sensory test that reordered its priorities was published in 2026. That is a reminder that a widely repeated brewing rule can rest on a mechanism rather than on a demonstration, and that the demonstration, when it arrives, may not rank the variables the way the mechanism implied.
The practical answer
Water composition genuinely changes how coffee tastes, and the variable that matters most is alkalinity rather than hardness. High bicarbonate buffers coffee's acids, making the cup flatter and more bitter, and the effect was detectable at small increases in a controlled sensory study, while calcium variation within realistic ranges made little difference. Pure water with nothing in it is not the answer either, since dissolved cations participate in extraction. Practically: filter out chlorine, avoid very alkaline water if you like bright coffee, and treat water as one variable behind grind, ratio and freshness rather than the first thing to fix. See also our article on espresso extraction and our article on coffee acidity.