Case Study 1 — Danny's Coffee Water Project

The fluorescent lights in the chemistry lab at the food science program had been buzzing for three weeks when the maintenance work-order finally cleared, and on the first morning they were quiet, Danny brought in his coffee.

He'd been thinking about this for months. The food science program ran a coffee club — fifteen students, mostly sophomores and juniors, who gathered every Thursday to taste coffee that one of them brought. Last week's offering, a Kenyan AA from a small Brooklyn roaster, had been rapturously received. Danny had bought the same beans, taken them home, brewed them on the same V60 dripper with the same paper filter and the same 1:16 ratio, and produced a cup that tasted flat. Not bad. Just flat. The brightness everyone had been talking about — the blackcurrant, the sparkle — wasn't there.

He'd asked the roaster's owner, who'd asked him about his water. Tap water? Danny had said. Filtered, but yeah.

That's your problem, the owner had said. Test your TDS. Bet you're high.

Danny had ordered a TDS meter that night. It arrived three days later. His apartment tap, post-Brita-filter, read 312 ppm.

The Specialty Coffee Association recommends 75–250 ppm. He was almost 25% over the upper bound. And the kind of dissolved solid mattered: most of his apartment's TDS, given Chicago's water profile, was probably calcium and magnesium carbonates — bicarbonates that buffered against the bright acids in the coffee, smoothing them out into nothing. The cup didn't taste sour because it didn't taste like anything.

So Danny started a project. He had a chemistry-teaching budget of about $40, a kitchen scale, a ream of pH strips, and a thesis: that he could systematically improve his daily coffee by understanding the water.

Phase 1: Mapping the problem.

He measured TDS at four points in his kitchen and the lab, on a Sunday morning when nothing else was running.

Two interesting observations. First, the Brita removed almost no dissolved minerals. Activated-carbon filtration handles chlorine, taste-and-odor compounds, and some heavier metals; it does very little for calcium and magnesium hardness. Danny had been telling himself for two years that "filtered" tap water was fine for coffee. It was fine for drinking. It was not fine for brewing.

Second, the difference between RO water and distilled water was negligible (and within the meter's noise floor). Either would work as a starting point for his experiments.

He brewed three cups of the Kenyan AA from each water source. Same beans, grinder, brew method, ratio, water temperature, all carefully controlled. He cupped them blind with two friends.

This was the result the SCA's recommendation predicted. Both extremes — too many minerals, too few — gave bad cups. The minerals weren't just neutral coolant; they were participants in the extraction.

Phase 2: Reading the literature.

Danny went to the library. He had access to a paper that the coffee community had been quoting for years: Hendon, Colonna-Dashwood, & Colonna-Dashwood, "The Role of Dissolved Cations in Coffee Extraction," published in Journal of Agricultural and Food Chemistry in 2014. The paper argued that the extraction of certain coffee compounds — particularly the citrate-bound molecules that contribute to brightness — was substantially enhanced by the presence of magnesium ions, while bicarbonate ions buffered against acidity and damped flavor.

Translation: he wanted moderate magnesium, low bicarbonate. The 2014 paper had even given a recipe — a mix of magnesium sulfate and sodium bicarbonate dissolved in distilled water at specific concentrations, providing a reproducible "ideal brewing water" for laboratory cuppings.

This appealed to Danny's instinct for control. He bought:

• 500 g pure magnesium sulfate (Epsom salt, food-grade — about $4)
• 500 g sodium bicarbonate (baking soda — about $2)
• A precision scale that read to 0.01 g (already had one)

He made a concentrate following the published recipe: a "buffer" solution and a "magnesium" solution, designed to be dripped at fixed ratios into a gallon of distilled water to produce the target water composition.

The first cup he brewed with his homemade water tasted transformed.

It wasn't a different coffee. It was the same coffee, with the dial turned. Suddenly the blackcurrant the roaster had described showed up. The acidity sparkled. The body was present. The cup he'd been brewing in his apartment for months had been a flat, muffled version of what the bean was capable of.

He cupped it with the same two friends. They both noticed. One of them, Han, said: "Wait — what changed?"

Danny told her, somewhat smugly. Han is a sensible person who doesn't drink the Kool-Aid easily. She made him brew her a tap-filtered version next to his lab-water version, with no labels, then taste them himself. He could pick the lab-water cup correctly four out of four times.

This is, in chemistry-class terms, not yet a result. Four trials is far below the threshold for blind-tasting confidence. But it was enough to convince Danny he was onto something real, and enough to motivate the next phase.

Phase 3: Diminishing returns and the 80% solution.

Danny took the project to coffee club. He laid out three cups, blind, prepared identically except for water:

• Cup A: tap water through his Brita
• Cup B: tap water through a higher-quality water-softener pitcher (a friend's, borrowed)
• Cup C: lab-water (his magnesium-bicarbonate recipe)

Eleven of the fifteen members came. He had them rank the cups for "brightness," "body," and "overall preference," anonymously, on index cards.

The results were boring and important. Cup C scored highest on brightness by a clear margin. Cup B scored highest on body, just barely. Cup A scored lowest across the board. But on overall preference, Cup B narrowly beat Cup C (5 to 4, with two ties), and only one person preferred Cup A.

This is what Danny had been suspecting. The lab-water recipe was technically the best — most extractive, brightest, most accurate to the bean — but it was also slightly too bright for some palates, slightly thin in body. The water-softener pitcher had produced a water with intermediate hardness, partially bicarbonate-reduced, which most tasters preferred for the actual cup.

The lesson was about diminishing returns. The first 80% of the improvement came from reducing total hardness and removing bicarbonate. The last 20% — getting to the laboratory ideal — required custom-mixed water and most people couldn't even tell, or even preferred the slightly less optimized cup.

Danny had become slightly humbled by his own data. He'd been ready to evangelize his recipe to the entire club. Instead, he wrote up his findings as: "For most people most of the time, a charcoal filter on the tap is the 80% solution. If you want to go further, pursue softening — reducing hardness reduces bicarbonate, which is the main suppressor of acidity. The custom-water route is real but yields diminishing returns and is taste-dependent."

Phase 4: The dissertation question.

Three months after the project began, Danny is now considering writing his thesis on it. The questions he wants to answer:

1. Is the magnesium-vs-calcium effect real, or is it a magnesium-vs-everything effect? The 2014 paper used magnesium sulfate. What if he used calcium chloride at the same concentration? Would the cup taste the same, or worse?

2. How does brew method interact with water? He's only tested pour-over. Would the same water optimization help espresso (where the much shorter contact time changes everything)? French press (where the long contact time might desensitize the cup to water composition)?

3. Can he develop a region-specific recommendation? Chicago's tap water profile is very different from Phoenix's, which is different from Atlanta's. A "use this water" recipe that's universal is less useful than a "given your local water, here's how to adjust" framework.

He's still working on it. The next coffee club meeting is Thursday.

📊 Analysis prompt:

1. Danny's TDS measurements found that activated-carbon filtration (Brita) removed effectively no dissolved minerals from his Chicago tap. What kind of filtration technology would substantially reduce hardness, and why?

2. Apply Henry's law to a sealed bottle of cold-brew coffee that's been refrigerating for 24 hours. The cold brew was prepared in a sealed jar with no special carbonation. What would Henry's law predict about CO₂ retention vs. a hot-brewed coffee left at room temperature for the same period? (Hint: think about the source of any CO₂ in each case.)

3. Danny's coffee-club tasting found that Cup B (intermediate water hardness) was overall preferred to Cup C (lab-water optimal). Translate this finding from a research-design perspective: what does "preferred" measure that "best extraction" doesn't, and how would you adjudicate between the two as a roaster trying to give brewing recommendations to your customers?