Extraction yield and TDS

What extraction is#

Brewing moves material from the ground coffee into water. Some of it dissolves. A coffee science textbook and a magazine article describe oil droplets and fragments of the cell structure being carried along as well in espresso.^11,14 According to a 2021 model study, only part of a roasted bean can dissolve: typically about 30%, or about one third of its mass.³

A coffee science textbook describes how compounds dissolve at different speeds. Very soluble compounds – caffeine, sugars, organic acids – come out in the first seconds of a brew.⁴ Less soluble bitter and astringent compounds keep coming out over the whole brew.⁴ In the textbook’s account, the balance of the cup therefore shifts as extraction goes on: from sweet and acidic towards bitter, harsh and astringent.⁴ The textbook gives this as the reason why “how much was extracted” is used as a quality measure.⁴

Two numbers: strength and yield#

TDS (total dissolved solids) is the strength of the brew. It is the amount of dissolved coffee in the beverage, expressed in per cent (or parts per million).⁸ It is usually estimated with a coffee refractometer, a small instrument that reads a drop of brew.⁸ A filter coffee at 1.30% TDS holds 1.30 g of dissolved coffee in every 100 g of beverage.

Extraction yield (also called percent extraction, PE) is the share by mass of the ground coffee that has dissolved into the brew.^2,8 It describes how much of the grounds has dissolved, not how strong the cup is.

The two are different things. According to a coffee science textbook, a strong coffee can come from little water and a lot of coffee, or the reverse.⁴ In either case, it may not taste right.⁴ In the textbook’s account, more coffee per litre at the same yield gives a stronger brew, not a more bitter one.⁴ The stronger brew simply contains more solids.⁴ Sensory research on drip coffee reports that, in general, bitterness rises with strength as well (see Contested).⁹

Ratios. Three masses are involved: the dose of dry coffee, the water poured in and the beverage collected.

• The brew ratio (brewing ratio) relates water to coffee. Researchers write it as water mass ÷ coffee mass.^2,3 The Specialty Coffee Association (SCA) home-brewer standard gives it in grams of coffee per kilogram of fresh water, with 55 g/kg as its test ratio.⁸
• In this entry, the beverage ratio means beverage mass ÷ dose: what is in the cup, not what went into the brewer. It is the ratio used in the extraction-yield formula below.^1,6 One espresso study gives its recipe this way: 20 g of coffee to 40 g of beverage.⁷
• The difference between the two is mostly retained liquid: the spent grounds hold back brew. In a 2023 drip-coffee study, this “liquid retention ratio” was about 2.1 g per gram of coffee.² In one set of immersion brews it averaged 2.48 ± 0.19.³

The core equation#

Extraction yield is calculated from the TDS, the mass of beverage and the dose:^1,6

EY (%) = TDS (%) × beverage mass (g) ÷ dose (g)

The worked examples below are our own arithmetic. They assume a filtered sample (dissolved solids only), a dose weighed dry, and the beverage weighed in the cup.

In the filter example, the 896 g of beverage is not measured but estimated. The grounds retain about 2.1 g per gram of coffee.² Taking this from 1,000 g of water leaves about 885 g of water, which carries 1.3% dissolved coffee. Weighing the beverage avoids this estimate.

Without a beverage weight, yield can be estimated from the water and the retention ratio. For drip brewing, the 2023 chart paper uses the formula below:

PE = TDS ÷ (1 − TDS) × (R_brew − R_abs)

Here R_brew is water mass ÷ coffee mass, and R_abs ≈ 2.1 is the retained liquid.²

On the chart, each brew ratio marks a diagonal line.² By this formula, the line is nearly straight and runs through the origin (our own arithmetic). So at a fixed ratio, a higher TDS always means a higher yield.

Measuring TDS#

Refractometer. A coffee refractometer estimates TDS from the refractive index of a drop of brew: how strongly the liquid bends light.⁸ Research labs cool the sample to room temperature before measuring.^1,3 One immersion study zeroed the instrument with distilled water.³ For espresso, labs pass the brew through a 0.45 µm filter first.^6,7

The SCA home-brewer standard (SCA 310-2021) specifies a coffee refractometer for the brew strength of every sample.⁸ Its separate test checks how evenly the grounds are extracted.⁸ That test requires at least three readings, averaged, with the refractometer well rinsed with distilled water.⁸

Oven drying is the reference method: a weighed sample of brew is dried to constant weight, and the residue is weighed.⁵ A 2013 comparison of nine brew methods dried its samples at 105 °C.⁵ Oven drying measures total solids, or dry matter, not only dissolved solids.^4,5 Refractometer readings are checked against oven drying¹ or against solutions of known concentration.³

Water “TDS” is not coffee TDS. According to a coffee science textbook, the handheld “TDS meters” used for water measure electrical conductivity.¹² They convert it with a factor that depends on the minerals present, so the result is only a rough estimate.¹² Coffee TDS from a refractometer is a different measurement.

The brewing control chart#

The control chart puts the two numbers on one page. Extraction yield runs along one axis, TDS along the other, and diagonal lines mark brew ratios.² Every brew is a point on it. According to an SCA magazine article by the chart’s researchers, all its quantitative relations follow from conservation of mass.⁹ In their account, the sensory zones drawn on it are what set it apart.⁹

History. The chart was developed in the 1950s.² In 1957, Ernest E. Lockhart of the Coffee Brewing Institute published The Soluble Solids in Beverage Coffee as an Index to Cup Quality.^1,2 The publication proposed that measuring TDS and calculating extraction could predict flavour and quality.^1,10 It identified TDS near 1.25% and extraction near 20% as “ideal”.³ The chart was later carried in the Coffee Brewing Handbook of the Specialty Coffee Association of America.^2,4 Studies from 2020 and 2023 describe it as the basis of vocational training in the coffee industry.^1,2

The classic zones. The classic chart is divided into nine regions. Along the TDS axis they are labelled “strong” and “weak”.¹ Along the extraction axis, the flavour labels are reduced to “under-developed” and “bitter”, usually called under- and over-extracted.^1,4,9 The central region – 1.15–1.35% TDS and 18–22% extraction – is the “ideal”, also known as the SCA Golden Cup Standard.¹ A coffee science textbook describes brews below 18% as under-extracted (too sweet and acidic) and brews above 22% as over-extracted (bitter and astringent).⁴

Sensory research from 2020. In studies published from 2020, a University of California, Davis (UC Davis) group brewed drip coffee to fixed points across the chart.^1,2 The group tested the brews with trained panels and consumers.^1,2 The findings:

• In a 2020 study, TDS had the greatest impact on the sensory profile; extraction yield had a lesser but noticeable one.¹
• According to the group’s SCA magazine article, bitterness in general rose with both TDS and extraction, as the classic chart implies (see Contested).⁹
• The same article reports that sourness rose with TDS but fell with extraction, peaking at high TDS and low yield.⁹ On the 2023 chart, sour, citrus and red-fruit notes sit in that corner; bitterness, astringency and roasted notes at high TDS and high yield.²
• According to two SCA articles, sweetness was highest at low TDS and low yield.^9,10
• In the 2023 study, consumers split into two groups. The larger (57%) liked low-TDS coffees best, with liking peaking at 18% extraction. The other (43%) liked two different kinds of coffee equally: low yield with medium to high TDS, and high yield with medium TDS.²

The authors call the classic “ideal” point misleading and possibly inaccurate. They propose a new Sensory and Consumer control chart to replace the classic one.² The new chart shows only drip coffee; it does not show the effects of variety, origin, roast or brewing variables.²

Plotting a brew on the chart

Enter your dose, the weight of the beverage (or the water) and the TDS you measured. The tool calculates the extraction yield with the formulas above and places your brew on the chart.

What changes extraction yield#

Extraction yield is an outcome, and many settings feed into it. According to a coffee science textbook, these settings interact: a finer grind both speeds extraction and lengthens flow time, so the effects multiply.⁴

• Grind. According to the textbook, finer particles expose more surface and extract more efficiently; too fine a grind can block flow.⁴ In a 2020 espresso study, measured yield peaked at an intermediate grind setting and dropped again at the finest settings.¹³ In a 2020 capsule study, a larger share of fines (particles below 100 µm) lengthened the extraction time.⁶
• Time and volume. In one espresso study, shots of under 10–15 s already reached 17–18% extraction.⁷ The best flavour for that coffee (one panellist) came at about 30 s and 19–20%.⁷
• Ratio. In percolation, water flows through a bed of grounds, as in drip brewing. According to the textbook, the balance in the cup keeps changing as more water passes through the bed.⁴ In full immersion, all the water steeps with all the coffee. In a 2021 immersion study, the brew approached an equilibrium: changing the ratio changed the TDS, but barely the yield.³
• Temperature. According to the textbook, solubility generally rises with temperature, and hotter water favours less polar, bitter compounds (polar compounds mix readily with water).⁴ Its sensory effect at a fixed TDS and yield was small in one drip study (see Contested).¹
• Water. According to the textbook, its hardness, acidity and cation composition can directly influence the sensory result of the extract.⁴ Cations are the positively charged mineral ions in water. The water entry covers this.
• Freshness. A coffee science textbook advises letting freshly roasted beans stabilise while they release gas.⁴ The freshness entry covers how this affects brewing.

Drip and immersion#

Drip (percolation). Water keeps passing through the bed. According to a coffee science textbook, the composition of the brew changes as long as it does.⁴ Beverage mass is easy to weigh, and the core equation applies directly.¹ In a 2023 drip study, the grounds kept about 2.1 g of liquid per gram, which never reaches the cup.²

Full immersion. All the water sits with all the coffee. In a 2021 model, the dissolved coffee spreads between liquid and grounds until it reaches equilibrium, a steady balance.³ The same study’s model and experiments put the yield at equilibrium at about 21% over a wide range of brew ratios.³ TDS was roughly inversely proportional to the brew ratio: a higher ratio gave a weaker brew.³ The authors conclude that immersion gives precise control of strength through the ratio but little control of yield.³ On the control chart, changing the ratio of an immersion brew moves it up or down, not sideways.³

That has a consequence for the arithmetic. The spent grounds of an immersion brew hold retained brew with dissolved solids, and the equilibrium calculation counts this coffee as extracted.³ Calculating yield from the beverage poured off therefore gives a lower number than the true extraction (see Deep dive).