Filter Coffee: Numbers, Myths, and the Paper Filter Effect
By Yalcin Akpinar · Food chemist, over 15 years in the flavour industry
Filter coffee is the most widely drunk and the most poorly documented brewing method of all. Almost every figure that circulates about it – 4 to 8 minutes, 1.15 to 1.35 per cent, the Golden Cup Standard – is either no longer correct or was never correct in the way it gets passed on. Here they are, with dates and sources.
First things first: none of this tells you how your coffee should taste. Numbers make a result repeatable, not right. If a coffee outside every target range tastes better to you, that’s not a mistake – it means you’ve found your range.
What is filter coffee, anyway?
Filter coffee means: water runs through a layer of ground coffee, and the finished coffee is separated from the grounds in the process. Technically, this is percolation – the water moves through the coffee bed and keeps meeting fresh material along the way. The opposite approach is immersion: in a French Press or a specialty coffee cupping, the coffee sits in the water and is only roughly separated at the end.
This difference isn’t merely academic. It decides how many fine particles and how much oil end up in the cup – and with that, body, clarity and, as you’ll see further down, blood lipids.
Who invented the coffee filter?
Melitta Bentz, 1908, in Dresden. That is well documented, yet almost everywhere it gets told wrongly.
On 20 June 1908 she applied for a utility model, described as a “coffee filter with a base that is domed on the underside and provided with a recess, and with obliquely directed flow holes”. It was registered on 8 July 1908, on page 1145 of the Patent Gazette of the Imperial Patent Office in Berlin. On 15 December 1908 the firm “M. Bentz” at Marschallstrasse 31 was entered in the commercial register – with start-up capital of 72 pfennigs.
Two details are regularly twisted: it was not a patent but a utility model. And it was not the 19th century but 1908. The dates come from the German Patent and Trade Mark Office.
Where do the filter coffee figures come from?
From the 1950s – and from a single research programme. Around 1952, the US National Coffee Association founded the Coffee Brewing Institute. Its first director was Ernest E. Lockhart, a food scientist who had earned his doctorate at MIT in 1939. In 1957 he published The Soluble Solids in Beverage Coffee as an Index to Cup Quality – the origin of the Brewing Control Chart that still hangs on the wall in every café today.
You often read that the chart was “developed at MIT”. The sources don’t support that: MIT was where Lockhart earned his doctorate; the coffee research took place at the Coffee Brewing Institute. We haven’t found any source that documents the research location.
What’s remarkable is how thin the basis is. A research group at the University of California, Davis re-examined the chart in 2023 and writes that neither the chart itself nor the source material makes it clear which people actually considered the coffee in the middle of the chart ideal. Their verdict on the “ideal” coffee marked there: misleading and possibly simply inaccurate.
Is the Golden Cup Standard still valid?
Not as a numbered standard. And that is more than hair-splitting.
Animation 1 · Strength (vertical) versus extraction (horizontal)
Both axes show percentages, but they measure different things. To the right: how much of the ground coffee has passed into the water (extraction). Upwards: how much of that ends up in the cup – in other words, the strength.
Schematic, not a measured value: both ranges come from the texts of standards, not from measurements. The diagonals of the original chart (brew ratio) have been left out. Sources: SCA “Coffee Standards, Revised 2018”, Section 5.1 · SCA-310-2021 “Home Coffee Brewers”: strength 1.15–1.55%, no requirement for extraction.
The well-known figures – 18 to 22 per cent extraction yield, 1.15 to 1.35 per cent dissolved solids – come from the SCA document Coffee Standards, Revised 2018, Section 5.1. This document is still on the SCA website, but it no longer appears in the current standards catalogue and has no SCA number.
Today’s catalogue contains no standard for brewing parameters. The published standards cover cupping (SCA-102 to SCA-105), equipment (SCA-310 for home brewers, SCA-320 for grinders, SCA-350 for espresso machines), training rooms and the qualification of examiners. A brewing parameter standard is not among them – nor is one in development.
In all honesty: we have not found an explicit statement from the SCA that the Golden Cup Standard has been withdrawn. All that is documented is that it is missing from the current catalogue. That conclusion is drawn by this article, not by the SCA.
It gets curious when you compare the numbers: the range against which the SCA certifies home brewers today is 1.15 to 1.55 per cent – considerably wider at the top than the classic 1.35. So anyone who cites “the SCA value” needs to say which one they mean.
Which brew temperature and which ratio?
Animation 2 · What the equipment standard prescribes – and what it doesn’t
The standard measures the brew temperature in the wet ground coffee, using three floating probes, not in the water tank. The often-quoted “4 to 8 minutes” come from the 2017 SCA certification programme; the 2021 equipment standard no longer says anything about it.
Source: SCA Standard 310-2021 “Home Coffee Brewers”. This is an industry association standard for equipment, not a study – and not a rule for your kitchen.
The current equipment standard SCA-310-2021 is specific on this point. A certified machine must be between 90 and 96 °C no later than the point at which a third of the water has been sprayed on – measured in the coffee bed itself, with three floating probes, not in the tank. As its standard ratio, the standard works with 55 grams of coffee per kilogram of water, which is around 17 grams per 300 millilitres.
For the carafe afterwards: 80 to 85 °C over the first 30 minutes, never above. If you’ve ever left coffee standing on a warming plate, you know the rest.
How long should filter coffee brew?
As long as it needs. There is no valid guideline for it – which surprises most people.
The famous four to eight minutes come from the requirements of the SCA certification programme for home brewers from 2017. It stated, word for word, that the contact time had to be “more than 4 minutes but less than 8 minutes”. The successor standard, SCA-310-2021, no longer contains any brew time requirement. The operating time is still measured, but only to determine when the temperature measurement is taken.
That makes sense, too: with percolation, the brew time is a result of grind size, filter and pouring technique – not a dial you can turn directly. How to use it in practice is explained in our V60 guide.
Is filter coffee healthier than unfiltered coffee?
On one point there really is a documented difference, and it has to do with the paper.
Coffee contains two diterpenes, cafestol and kahweol. They are dissolved by the hot water but are largely retained by the paper filter. In controlled studies, cafestol raised cholesterol levels measurably: in a randomised study from 1997, total cholesterol rose by around 31 milligrams per decilitre with 61 to 64 milligrams of cafestol a day over four weeks. A 2012 meta-analysis of twelve controlled studies with just over 1,000 participants found, on average, just over 8 milligrams per decilitre more total cholesterol – with the larger increase for unfiltered coffee.
Animation 3 · What the paper filter holds back
Cafestol is a fat-like substance from the bean that raises cholesterol. All bars start at zero, so the length comparison holds. Where the thin line is missing, no spread has been published for that method. Only 3 to 11 samples were measured per method – the bars are snapshots of a few cups, not fixed values. Espresso is deliberately left out: four samples ranged between 36 and 2,447 mg/l, which doesn’t allow a meaningful bar to be drawn.
Spread in detail: coffee machine 24–444 mg/l (11 samples) · paper filter 4–24 mg/l (5 samples) · French Press 87, percolator 91 · cloth filter only one series of measurements · boiled coffee, unfiltered, no published spread. Median = the middle value when all measurements are ranked by size. Source: Orrje et al., Nutrition, Metabolism and Cardiovascular Diseases 35 (2025).
Measured values, not a model calculation – but only three to eleven samples per method.
A Swedish series of measurements from 2025 shows particularly clearly how big the difference between brewing methods is. Scandinavian boiled coffee had a median of 939 milligrams of cafestol per litre, home-made paper-filter coffee 12. The same boiled coffee poured through a cloth filter dropped to 28.
At the level of whole lifetimes, a Norwegian cohort study took a look in 2020: 508,747 people, followed for an average of 20 years. Filter coffee was associated with lower all-cause mortality than no coffee at all; unfiltered coffee fared worse than filter coffee.
What these figures don’t say
- Unfiltered coffee has not been shown to be harmful. In most groups of the Norwegian study it was statistically no worse than no coffee at all – the confidence intervals include the neutral value.
- These are observational data. There is no randomised study comparing filtered with unfiltered coffee using heart attack or mortality as the endpoint.
- Filter coffee isn’t at zero either. The Swedish series found 12 milligrams per litre – little, but not nothing.
- Espresso varies enormously. Individual samples were above 2,400 milligrams per litre. On this data, a blanket assessment of espresso isn’t possible.
How do official bodies see it? Norway is the only country where we found an explicit recommendation: the Norwegian Directorate of Health writes that one to four cups of filtered coffee a day can be part of a healthy diet for adults, and in the same breath defines boiled coffee and press-pot coffee as unfiltered. The Nordic Nutrition Recommendations 2023 advise against a high intake of boiled coffee.
For the European Food Safety Authority, Germany’s Federal Institute for Risk Assessment and the relevant Swiss authorities, we found no statement on coffee diterpenes – there, the focus is exclusively on caffeine. That doesn’t mean none exists; we simply didn’t find one. And of course this section is no substitute for talking to your doctor.
Which grind size and which beans?
With grind size, the direction matters more than the number. For cone filters, the SCA equipment standard gives a target range of 425 to 850 micrometres, and 589 to 1,168 micrometres for flat-bottom filters – roughly speaking, about like coarse sea salt, much coarser than for espresso.
In practice you don’t need a sieve analysis for this, just two observations: if the coffee runs through very quickly and tastes thin and sour, grind finer. If the water sits for a long time and the coffee turns harsh and dry, grind coarser. What you need for this is a grinder with repeatable settings – you’ll find a selection among our coffee grinders.
EstimateTwo statements in this article don’t come from a measurement: the comparison with coarse sea salt, and the rule of thumb that boiled water is in the 90 to 96 °C range after about a minute. Both are values from experience in our courses – how quickly your water cools depends on the kettle, the amount and the room temperature.
When it comes to beans, there is no right answer. Lighter roasts show more fruit and acidity in the filter, darker ones more chocolate and body. How a given Arabica variety tastes also depends on origin and processing – and, in the end, on you. If you want to know what the score on the bag tells you about this: Specialty Coffee.
Which water?
This is where the sources get messy, and we would rather say so than hand you a number.
The 2018 SCA document gives target values of 150 milligrams of dissolved solids per litre, calcium hardness of around 68 milligrams per litre, alkalinity close to 40 milligrams per litre and a pH of around 7.0. The 2021 equipment standard SCA-310 refers to “the current SCA water standard” and in doing so gives different values: 50 to 175 ppm calcium hardness and pH 6 to 8.
There is currently no water standard at all in the numbered SCA catalogue, and two SCA documents give different figures. Which ones apply today cannot be answered unambiguously from public SCA sources. What remains reliable in practice: no chlorine, not rock-hard, not distilled.
Frequently asked questions about filter coffee
What is filter coffee?
Filter coffee is coffee made by letting hot water run through a layer of ground coffee, with the finished coffee separated from the grounds by a filter along the way. That sets it apart from full-immersion methods such as the French Press, where the coffee sits in the water and is only roughly strained off.
What is the right brew temperature for filter coffee?
The current SCA equipment standard SCA-310, from 2021, requires 90 to 96 °C, measured in the coffee bed itself. In practice this means: bring the water to the boil and let it stand for about a minute, and you’ll be in this range.
How long should filter coffee brew?
There is no binding guideline. The often-quoted four to eight minutes come from the 2017 SCA certification programme; the current standard SCA-310-2021 no longer includes a brew time. The brew time is a result of grind size and filter, not a goal in itself.
What is the right coffee-to-water ratio?
55 grams of coffee per litre of water has become the established starting point – that’s around 17 grams per 300 millilitres. This figure is also in the current SCA equipment standard. Whether you prefer it stronger or milder is not something it answers.
Is the Golden Cup Standard still valid?
The figures of 18 to 22 per cent extraction and 1.15 to 1.35 per cent dissolved solids come from the SCA document “Coffee Standards, Revised 2018”. Neither a Golden Cup Standard nor a standard for brewing parameters appears in the SCA’s current numbered standards catalogue. However, the SCA has not published a formal withdrawal.
Who invented the coffee filter?
Melitta Bentz from Dresden. On 20 June 1908 she applied for a utility model, which was entered in the Patent Gazette of the Imperial Patent Office on 8 July 1908. It was expressly a utility model, not a patent – and it was 1908, not the 19th century.
Is filter coffee healthier than unfiltered coffee?
When it comes to cholesterol levels, there is a documented difference: a paper filter largely holds back the diterpenes cafestol and kahweol, which raised LDL levels in controlled studies. A Norwegian cohort study of just over 500,000 people found lower mortality with filter coffee than with unfiltered coffee over twenty years. These are observational data, not proof of causation – and in most groups, unfiltered coffee fared no worse than no coffee at all.
What grind size does filter coffee need?
Much coarser than espresso, roughly like coarse sea salt. For cone filters, the SCA equipment standard gives a target range of 425 to 850 micrometres. The direction matters more than the number: if the coffee runs through too quickly and tastes thin, grind finer. If the water sits and the coffee turns harsh, grind coarser.
What now?
The most useful first step isn’t buying a refractometer, but brewing two carafes side by side: one at 55 grams per litre, one at 65. The difference is big enough that you’ll notice it straight away – and afterwards you’ll know where your range lies, rather than where the one from 1957 lay.
You’ll find it step by step in our V60 guide. If you want to do the same with espresso: Espresso dial-in. And if you’d rather go through it once with someone guiding you: the workshops in Bern and Zurich. We are United by Coffee – whatever the brew method.
Sources and context
- Invention of the coffee filter: German Patent and Trade Mark Office, article on Melitta Bentz (application 20 June 1908, registration 8 July 1908, page 1145 of the Patent Gazette, commercial register entry 15 December 1908).
- Golden Cup Standard: Specialty Coffee Association, Coffee Standards, Revised 2018, Section 5.1. That neither a brewing parameter standard nor a Golden Cup Standard is listed in the current standards catalogue is a negative finding from the SCA’s official list of standards.
- Brew temperature, ratio, grind size, holding temperature, absence of a brew time requirement: SCA Standard 310-2021, Home Coffee Brewers: Specifications and Test Methods.
- Contact time “more than 4, less than 8 minutes”: SCA, Minimum Certification Requirements for Coffee Brewers, Certified Home Brewer Program, 2017 – a superseded version of the programme.
- Coffee Brewing Institute, Lockhart and the 1957 publication: Emma Sage, The Coffee Brewing Institute, SCA News 2013; Frost, Batali, Guinard, Ristenpart, Towards a New Brewing Chart, 25 Magazine No. 13.
- Criticism of the classic chart: Guinard et al., A new Coffee Brewing Control Chart relating sensory properties and consumer liking to brew strength, extraction yield, and brew ratio, Journal of Food Science 88(5), 2023.
- Diterpenes and cholesterol: Urgert & Katan, Annual Review of Nutrition 17, 1997; Urgert et al., American Journal of Clinical Nutrition 65, 1997; Cai et al., European Journal of Clinical Nutrition 66, 2012.
- Cafestol content by brewing method: Orrje et al., Nutrition, Metabolism and Cardiovascular Diseases 35, 2025.
- Cohort study: Tverdal et al., Coffee consumption and mortality from cardiovascular diseases and total mortality, European Journal of Preventive Cardiology 27(18), 2020.
- Official guidance: Helsedirektoratet (Norway), Kostrådene, as of 15 August 2024; Nordic Nutrition Recommendations 2023, Beverages chapter.
- We have no statement from EFSA, BfR or Swiss authorities on coffee diterpenes; the text says so explicitly.
Yalcin Akpinar
Co-founder, equipment and sensory
Graduate food chemist from the University of Hamburg, over 15 years in the international flavour industry, patents in flavour technology. With us he is responsible for roast profiles, bean selection and the choice of machines and grinders.
