There’s a cafe in Georgetown, Penang, designed by SK Leng, a man with a mild obsession for coffee. Under the espresso counter he’s got pipes running into empty upside down filter housings (to act as a small reservoir) and BWT bestmax PREMIUM ion-exchange filters stripping out calcium and adding magnesium to his water with testing points and connections for planned adaptations in several locations. His irrigation system is constantly changing; I’ve never visited this cafe without seeing some change to the way that he waters his coffee. To most people, water is H2O and that’s it. To health conscious GMO avoidant modern consumers, removing calcium and adding magnesium seems like the work of Dr. Frankenstein. Why would anyone want to do all this to pure, natural water?
Water ain’t just water
Water is almost never naturally just H2O. Depending on where you are in the world and what soils, lakes and sewers the water has been through, the water will have many other molecules mixed up with the H2O. The make-up of the water has a large bearing on how it extracts water from coffee. A 200 mL filter coffee is made up of around 2.5 g of coffee. The rest of the weight – something like 98.75% – comes from water. So, if we’ve spent lots of money getting the best equipment and acquiring the best beans, it makes sense that we should pay careful attention to the 98.75%.
It doesn’t take someone of Frankenstein’s brilliance to understand how to manipulate water, but it does involve some chemistry that can appear alchemical if you haven’t come to coffee via a science degree or the book Water for Coffee by Christopher Hendon and Maxwell Colonna-Dashwood. The book is less of a simple how-to guide and more of a extended lesson on water chemistry and its interaction with coffee and goes a long way towards making the creation of water seem more feasible.
How to brew beer water
I came to water manipulation via home brewers forums where men with various ‘b-shaped’ torsos would swap recipes for beer with names like ‘Firework Cream Ale’ and ‘Moose drool’. Along with the ingredients for the beer were specifications for the water that included ‘general hardness’, ‘carbonate hardness’, and ‘pH’, which are all important for coffee extraction. They also include ‘sulphate concentration’, ‘pitch rates’, ‘wort volumes’, ‘Lauter headspace’ and many complicated calculations, which either suggests that beer brewing is more complicated than coffee brewing or that people have spent more time obsessing over brewing their beer than their coffee. The importance of water for beer is well understood and, in one of my favourite online calculators (Brewer’s Friend water calculator), we can dial up various water profiles to match our water to a particular style of beer. Noticeably, these styles are usually associated with places (see screenshot below). This is because the water chemistry in different areas is different depending on the soils and pipes and treatment of the water on its way from cloud to conical cylindrical fermentation tank and different water lends itself to different styles of beer. It wasn’t that Londoners didn’t like the light beers of the Czech Republic, rather with higher levels of calcium in their tap water, the London brewers would have gotten astringent, green flavours from the beer if they’d tried to brew Pilsner. Instead they worked with what they had and brewed up world famous London porters and dark ales.
Just have a wee think about this in terms of coffee. A roaster in London will roast their coffee to taste good with the very hard water of London. Presuming that the London roastery uses London water to cup, they will have roasted their Ethiopian Wegida so that it has honey sweetness, lemonade acidity and notes of peach and all the cafes in London could achieve these notes if only they prepare the coffee well. Bruce, however, with his soft Melbournian water will probably find that his expensively imported coffee is tasting less peach, more sour apricot; and less lemonade, more lemon zest. London water supply has qualities that make coffee heavier, flatter and chalky and anyone roasting for this water will need to develop a roast profile so that it is less soluble, more sour and sharper so that it tastes good when extracted with this water. Bruce, however, gets a brew that is empty, sharp and sour no matter how hard he tries to vary his preparation method and is left confirmed in his suspicion that Melbournian coffee is the best on the planet.
If you’re getting your beans roasted in a different location, are they using the same water as you? Most speciality cafes use locally roasted beans. This doesn’t only make supply easier, but also makes it more likely that the coffee will taste the same as when it was on the roaster’s cupping table.
Although I do have a science background, I found the information available on beer brewing books confusing and wasn’t sure how much time to spend testing the effect of sulphate concentrations on my coffee brewing. This is partly because the beer brewing scene is full of well-meaning, but largely misinformed amateurs the length of whose moustaches were proportional to the fibs they told about water chemistry. Another thing that complicated my inquisition was that instead of measuring chemicals in parts per million or mg/L many test kits give a result in parts per million as a concentration of calcium carbonate, which is not helpful in the slightest and needs converting. This is an historical anomaly that has become the industry standard and explained well in the Water for Coffee Corrigendum. It wasn’t until the publication of this book that I was confident that my time spent experimenting with water composition wasn’t time spent piddling into the ocean.
To be clear: the best reason to read and understand this book is so that you can manipulate your water because the composition of the brew water makes a huge difference to the resulting flavour of your coffee.
Three main factors for coffee brew water quality
GH and KH
Right now, you’re probably not getting the best out of your coffee – even if you work in the cafe attached to the roastery. Water extracts coffee through three different processes: dissolution, hydrolysis and diffusion and the speed at which it does this is determined by the concentration of calcium and magnesium ions in the water; they give the water “pulling power” and are referred to as General Hardness [GH]. Water with low GH makes coffee that feels empty; water high GH gives us coffee that feels heavy. This aggressive ripping of coffee compounds out of the grinds and into the water would make the brew quite sour if it weren’t for another chemical that’s usually in our water: carbonate. The concentration of carbonate ions in the water is referred to as Carbonate Hardness [KH]. Water with low KH makes coffee that is sour; water with high KH is chalky. Think of GH as Special Forces units sent in to extract your coffee and KH as the diplomat to smooth things over.
A third factor that Hendon and Colonna-Dashwood mention is the pH, which is the concentration of hydrogen ions in relation to hydroxide ions. pH is not something that is focussed on other than to suggest that the ideal pH would be slightly basic (i.e. pH 7-8) in order to increase the number of redox reactions and as a possible counter to the slight acidity of coffee.
There you have it. What’s important for us coffee brewers in water is GH, KH and pH. And Hendon and Colonna-Dashwood go further and suggest a zone of ideal brewing that falls between 50 and 200 ppm for GH and 37 and 75 ppm for KH – the ideal level of GH is dependent on the level of KH, so check out the graph on page 108 for more precise numbers. Presumably, they made many samples of water and tested these with a coffee. We’re not told how the coffee was roasted or what preparation method they used, so we can’t say that this range works well for every style of coffee or preparation method. But we’ve tested out our coffee at the roastery and brewery using coffee roasted lightly in a Loring Smartroast and prepared using v60 (drip brewing) and Clever Drippers (immersion brewing) with water straight off the boil and also with cold brewed coffee in 150 L tanks (immersion brewing) over 12 hours with water at 25ºC. Using water with GH of 120 ppm, KH of 60 ppm and pH of 8, various coffees all tasted richer with more body and accentuated notes compared to our typically soft water. It mightn’t improve coffee roasted in a different style, but it works well for us.
How to know if a filter salesperson is bullshitting you
He is bullshitting you. 98.75% guaranteed.
You need to measure the GH, KH and pH of your supply water. In none of the cafes that I own do we need expensive filters. If your cafe is in a city with modern water treatment, you might not need a filter at all. I do recommend that all cafes use a carbon block filter. These are cheap and filter out various nasties that might be in the water and can act as a backstop should something go wrong upstream. If your supply water is soft (i.e. the GH and KH are low), adding more filters will probably make your coffee taste empty and sour. You only need filters to reduce your GH if it’s over 175 ppm and your KH if it’s over 75 ppm. If your GH is made up of calcium ions (it probably is), it is worthwhile stripping out the calcium and replacing it with magnesium ions to prevent limescale buildup on your expensive coffee equipment. Ask the filter salesperson what the GH and KH are in your area and if she starts talking about hardness (ppm) levels, ignoring GH and KH, or she gets out an Ionic Conductivity TDS meter to prove her point, she’s bullshitting you.
The solution that best suits your water is entirely geographically dependent. Some regions have naturally occurring soft (low GH) water and might only require a carbon block; other filtration methods would take even more of the naturally occurring minerals from the water, making it increasingly unfavourable. The problem, however, is that water composition varies dramatically from city to city, and depends on the minerals the surface water interacts with. p63, Water for Coffee.
Measuring the factors
You may be able to look up the GH, KH and pH of your water using the website of your local water supplier, but if you write to them and get brushed-off with a bland cut-and-paste response from some kid wasting his English degree in Bangalore, you need to measure it yourself.
There are various ways of measuring these, but the cheapest and most reliable are titration kits from a reefer store. Reefers are people with a mild obsession with coral and they care a lot about water. They hide in the back aisles of pet stores. Don’t approach them, they are easily frightened away. Just speak to a shop assistant. You can get kits for magnesium ion concentration, calcium ion concentration, magnesium and calcium combined (GH), KH and pH. Make sure that the pH kit tests at least in the range of pH 5-9.
Once you have your numbers you can measure up on the brewing zone chart (page 108 of Water for Coffee) to see if you need to add or subtract magnesium, calcium and carbonate from your water. If your GH and KH are too high, you’ll need to remove some magnesium, calcium and carbonate ions, before testing again to see how much magnesium and carbonate ions you need to add to get the GH and KH back into the zone. If your GH and KH are too low, you can skip the removal step and just add magnesium ions and carbonate ions.
How to brew coffee water
Essentially the process is this:
- Take a sample of your water
- Test the GH and KH
- Check your levels against the brewing zone chart
- If your GH or KH are too high, take some out
- If your GH or KH levels are too low put some in
- Test your pH and hope that it is between 7 and 8.
Here’s where things get expensive. The best options depends on your location. I recommend three: Reverse Osmosis [RO], ion-exchange filters or dilution with deionised water. The advantages of each are discussed at length in the book, but the simplest option would be to buy tanks of deionised water and to simply add a mix of ions in each time.
Calcium doesn’t add much to brew water that magnesium can’t (those hardworking chaps, Hendon and Colonna-Dashwood, published a paper showing as much in 2014). And calcium combined with carbonate leads to the limescale that chokes up espresso machines everywhere. So, to increase GH, just add magnesium chloride or magnesium sulphate.
Carbonate is easily obtainable at the supermarket as sodium bicarbonate, or baking powder, but you can also use potassium bicarbonate. So, to increase the KH, just add baking powder.
How much to add?
Good question! The answer depends on what you’re adding and how much water you’re treating. The equations for all of this are in the book, but adding 1 g of magnesium chloride to 1 litre of water increases the GH by 119 ppm. Add 1 g of sodium bicarbonate to 1 litre of water increases the KH by 726 ppm.
What about changing the pH?
Don’t unless you really need to. Changing the pH is possible and not too difficult, but is likely to change your KH, so you may use up all your titration kits up adjusting the KH and the pH levels. If your pH is anywhere between 7 and 8, that should be fine.
If you are going to meddle with the pH, get your safety googles and gloves on: it’s not for the faint-hearted or those wishing to remain aesthetically appealing. We can’t easily pre-determine the amount that pH will change when we add our chemicals. We just add a bit and then retest the pH, then add a bit more and retest the pH etc. Because this requires many measurements, I prefer to use a pH meter rather than to titrate each measurement.
To reduce the pH, add one drop (0.05 mL) of hydrochloric acid at a concentration of 0.01 g/L for every litre of water to reduce pH, then test again.
To increase the pH, add a 0.01 g piece of sodium hydroxide for every litre of water to increase pH, then test again.
And now test your KH again because that’s probably changed.
Want an easier way to brew up your brew water?
- Buy a tank of deionised water
- Dissolve in a glass:
- 1 g of magnesium chloride hexahydrate per litre of water in the tank
- 0.08 g of sodium bicarbonate per litre of water in the tank
- Add the dissolved minerals to the tank and mix.
- Brew coffee
The book Water for Coffee is essential reading for anyone serious about their coffee. It is complicated going – a friend who is both a coffee professional and who studied analytical chemistry still found the book hard going. However, the book acts as a chemistry primer, taking things right back to the beginning – Section 1 of Chapter 1 is titled ‘The birth of the Universe’ – and explaining all the concepts in depth. It just may take several readings to fully grasp all the concepts. Despite the complexity of the subject, it is a thoroughly readable book helped along by relevant and amusing interjections by Colonna-Dashwood. This is one of the best books on water out there and it happens to focus on coffee. Highly recommended.
Last I heard, SK Leng has been working in the back room of his cafe on a system that has two large pumps, a reverse osmosis system and three large stainless steel tanks up by the ceiling. Each tank contains enough water to supply 150% of the cafe’s daily supply. The reverse osmosis system creates waste water, which is pumped into a fourth tank that supplies the dishwashing sink, and mineral reduced water, which is pumped into the first tank up near the ceiling.
Once the first tank has filled up, the water gets pumped to the second tank where the appropriate salts will be mixed in by a member of staff. The third tank is for the coffee supply. Once it’s empty, the mineralised water from the second tank gets pumped through and the RO water from the first tank can be pumped into the second.
Further in the future, SK hopes to have an arduino bot mix the appropriate salts and drop them into the tank using a modified inkjet printer. But first, it’s all about water test strips followed by milligrams of magnesium chloride and sodium bicarbonate. And great tasting coffee.