What Mineral-Heavy Water Does to Oolong: Less Catechin, a Cloudier Cup
A 2024 study brewed the same Tieguanyin in four different waters and measured what came out. Mineral content moved the chemistry, not just the story a taster tells about it.
Water temperature gets the attention the brewing guide already codifies: 85 to 90C for a green-leaning oolong, closer to a full boil for a heavy roast. Mineral content gets none. That is a gap, not an oversight the science excuses. A 2024 study brewed the same Anxi Tieguanyin, both its light-scented and strong-scented styles, in four different waters at fixed temperature and time1, then measured the catechins, the aroma compounds, and the color that actually came out. The waters produced different tea. Not a different impression of the tea. Different measured chemistry.
Four waters, one Tieguanyin, and a measurable difference
Researchers at China's Tea Research Institute of the Chinese Academy of Agricultural Sciences, led by Ma and colleagues, brewed both Tieguanyin styles in four commercially available waters: a purified water (Wahaha), a mountain spring water (Tongsheng), a bottled mineral water (the Fiji brand, sold in China as Natural Waters of Viti), and a "natural water" (Nongfu Spring), a category between spring and mineral in China's own bottled-water standard. The four sat in two clear bands before a kettle was involved. The purified and mountain spring waters measured below pH 7 at room temperature; the mineral and natural waters measured above pH 7.5. Boiling pushed every sample's pH up by about 0.78 units, closing the gap slightly but not erasing it. Conductivity, a rough stand-in for total mineral load, told the same story: the mineral water measured 300 microsiemens per centimeter, the highest of the four by a wide margin, with calcium, magnesium, sodium, and bicarbonate all running highest in that same sample (Ma et al., Food Chemistry: X, 2024).
What the minerals did to the leaf's chemistry
For the light-scented Tieguanyin, the mountain spring water produced the highest sensory acceptability of the four; the Fiji-brand sample produced what the researchers' tasting panel recorded as a flat taste and a turbid aroma. The instrumented numbers back the panel up. Total epi-catechins in that same infusion measured 82.25 milligrams per liter, against 276.48 in the mountain spring water, 263.46 in the purified water, and 260.70 in the natural water. Epigallocatechin gallate, the single catechin most responsible for an oolong's astringent bite, told the sharpest version of the same story: 27.49 mg/L in it against 131 to 144 mg/L across the other three. Total aroma volatiles ran lowest in the purified water and highest in the natural water, with the mineral brand in between. Color moved with pH and conductivity too: its infusion measured the darkest of the four, a relationship the authors report as a strong negative correlation between lightness and mineral load (r = −0.978). The strong-scented Tieguanyin, roasted and lower in remaining catechin to begin with, repeated the catechin result almost exactly: the same bottled brand again produced the lowest total catechins, 85.79 mg/L against a high of 386.79 in the natural water. The aroma numbers were messier this time, the purified water actually gave the lowest total volatiles for this style, with the natural water still highest, and a fruitier top note replacing the light style's floral one. So the catechin suppression held across both styles at fixed temperature and fixed time, with the leaf held constant; the aroma effect did not track the same water sample as cleanly.
Why alkaline water breaks catechins down, not just dilutes them
The Tieguanyin study measured the outcome; a separate 2021 study on green tea measured the mechanism, and it is worth borrowing because catechins do not know which plant they came from. Cabrera and colleagues brewed the same tea in synthetic waters set to five hardness levels, 21, 42, 85, 169, and 338 parts per million as calcium carbonate equivalents, running from very soft to very hard, plus a deionized-water control. Total catechin content fell from 164.6 milligrams per cup in the very soft water to 67.5 in the very hard water, a 2.4-fold drop. Epigallocatechin gallate and epigallocatechin, the two catechins carrying three hydroxyl groups on their B-ring, fell 3.2-fold and 3.1-fold respectively, the steepest losses of any compound measured. Caffeine barely moved. Catechin, epicatechin, and epicatechin gallate, the compounds without that third hydroxyl group, fell only 1.4 to 1.7-fold (Cabrera et al., Molecules, 2021).
The mechanism the authors identify is not simple dilution. It is degradation: autoxidation followed by polymerization, a reaction that alkaline conditions accelerate, and mineral salts accelerate it further while also reducing how much catechin the water pulls out of the leaf in the first place. Harder water in the Tieguanyin study measured a higher pH after boiling, the same alkaline direction. This is why the effect should matter most at the lighter end of the category. A jade Tieguanyin or a high-mountain gaoshan still carries most of its catechin load into the pot, while a roasted yancha converted much of that catechin during processing before a kettle enters the picture, so the dark end has less left to lose to a mineral-heavy water. The same asymmetry the temperature research maps between the two ends of the category shows up again here, driven by a different variable.
The specific point where hardness turns the cup cloudy
Calcium does one more thing past a threshold, and it is a threshold, not a gradient. Oolong liquor carries organic acids, oxalic and tartaric among them, left over from the leaf. When dissolved calcium in the brewing water exceeds about 40 milligrams per liter, calcium forms insoluble precipitates with those acids directly in the cup, the mechanism behind a visibly cloudy infusion and the thin, chalky-tasting film sometimes called tea scum. The Tieguanyin study's mineral water, the sample with the highest calcium and magnesium load of the four, is exactly the one its own tasting panel flagged for a turbid aroma. This is a different failure mode from the catechin suppression above, one is a chemical breakdown, the other a physical precipitation, but both land on the same conclusion: past a specific mineral concentration, the water stops being a neutral carrier and starts editing the tea.
A number to work from, not to chase precisely
The Tea Association of the U.S.A.'s own brewing manual, issued jointly with the National Restaurant Association, states the threshold plainly: tests show hardness in excess of 200 parts per million can cause clouding in iced tea, and the manual names clarity, color, and taste as the three properties high mineral content threatens first. The same manual notes that iced tea accounts for 80 to 85 percent of American tea consumption, which is presumably why an industry association bothered writing the number down. Total dissolved solids and the conductivity the Tieguanyin study actually measured are related but not identical figures, so treat the comparison loosely: the mineral water's 300 microsiemens per centimeter reads high against that 200 ppm ceiling on any reasonable conversion, while the mountain spring and natural waters, the two the tasting panel preferred, read closer to it. None of these numbers is a recipe to replicate at the kettle. They are a ceiling to stay under, the way the brewing guide's temperature bands are a range, not a single correct degree.
The water least likely to blunt a jade oolong
Distilled water is not the answer either. Stripped of every mineral, it extracts differently again and tends to read flat, the same complaint the Tieguanyin panel logged for the over-mineralized end, for the opposite reason: there is nothing left in the water to carry the leaf's chemistry into the cup evenly. A basic carbon filter pitcher, run on ordinary tap water, removes chlorine taste and a meaningful share of temporary hardness (the calcium and magnesium bicarbonates that boiling itself partly breaks down) without demineralizing the water outright, which puts most municipal tap supplies closer to the range the panel preferred than either extreme. For a jade-style Tieguanyin or a high-mountain gaoshan, the water least likely to blunt the floral top notes is a moderately mineralized one, filtered tap or a "spring" or "natural" labeled bottled water rather than a heavily mineralized brand. A dark, charcoal-roasted yancha or dong ding, having already lost much of the catechin load a hard water would otherwise degrade, tolerates more variation in the water without the same penalty. That gap tracks the one temperature already opened between the two ends of the category. It just runs through the tap, not the kettle dial.