{ "id": "water-treatment", "url": "https://coffearium.org/entries/water-treatment/", "license": "CC-BY-4.0", "title": "Water treatment", "summary": "How filters, ion exchangers, reverse osmosis and added salts change brewing water – what each removes or adds, which suits which tap water, and how to keep a treatment system clean and working.", "domain": "extraction", "kind": "Concept", "stage": "brew", "topics": [ "equipment", "science" ], "topicSections": {}, "lang": "en", "status": "verified", "reviewer": "Coffearium verification pass", "lastVerified": "2026-10-20T00:00:00.000Z", "updated": "2026-10-20T00:00:00.000Z", "added": "2026-10-20T00:00:00.000Z", "brief": [ "Each treatment moves water in its own direction – a softener lowers hardness only, a decarboniser lowers hardness and alkalinity equally, reverse osmosis removes almost everything.1", "Activated carbon removes chlorine, chloramines and organic off-flavours; chlorine left in brewing water can react with coffee phenols to form medicinal-tasting chlorophenols.1,2", "A softener swaps calcium and magnesium for sodium, so alkalinity stays and the sodium content rises.1,7,8", "For espresso machines, the Specialty Coffee Association (SCA) advises a decarboniser bypass giving 40–70 ppm alkalinity, or, above 250 ppm input, removing at most 200 ppm.1", "Carbon filters and other point-of-use devices can host bacteria, so replacement and sanitising are part of any treatment.2,9,10" ], "quickFacts": [ { "label": "Removes taints", "value": "Activated carbon (chlorine, chloramines, organic matter)" }, { "label": "Lowers hardness only", "value": "Softener (Ca²⁺, Mg²⁺ → Na⁺ or K⁺)" }, { "label": "Lowers hardness and alkalinity", "value": "Decarboniser (Ca²⁺, Mg²⁺ → H⁺), with bypass" }, { "label": "Removes almost everything", "value": "Reverse osmosis, deioniser; then blend or remineralise" }, { "label": "Raises minerals", "value": "Hardener, alkaliser, carboniser, recipe salts" }, { "label": "Decarboniser for espresso (SCA)", "value": "40–70 ppm alkalinity out; remove at most 200 ppm" }, { "label": "By-products", "value": "Sodium (softener); CO₂ and lower pH (decarboniser)" } ], "aka": [ "Water filtration", "Water filter for coffee" ], "linkTerms": [ "water treatment", "water filter", "water filters", "water filtration", "treated water", "filter cartridge" ], "related": [ "water-for-coffee", "hardness-and-alkalinity", "measuring-water", "water-minerals-and-flavour", "espresso-machine", "espresso", "brew-methods" ], "contested": [ "Bacteria in carbon filters: in a 1995 study of rural groundwater supplies, effluent from point-of-use carbon filters had more heterotrophic bacteria after overnight standing, but fewer than the incoming water after a short flush, and the authors concluded that, apart from water that had stood overnight, the filters did not compromise water quality.9 A 2004 study of household tap filters found the highest bacterial counts in filtered water and suggested that biofilm in the devices may amplify bacteria.10", "Softened water for brewing: the SCAA handbook advises against it, citing slower flow through the coffee bed and over-extraction in Coffee Brewing Center trials;2 the SCA 2018 handbook lists the softener as one of the most common treatments and describes only its effect on hardness and alkalinity.1", "How much carbon dioxide a decarboniser adds to espresso: the SCA handbook and a textbook chapter by the same research group each give a worked example, with different alkalinity reductions and different estimates of the share added to the carbon dioxide in fresh coffee.1,16 No measurement of this effect was found for this entry." ], "corrections": [ { "date": "2026-10-20T00:00:00.000Z", "change": "The 1995 carbon-filter study's conclusion now keeps its exception for water that stood overnight; the 2004 study's biofilm explanation is given as a suggestion, as in its abstract.", "reason": "Both statements were stronger than the abstracts of the two studies." }, { "date": "2026-10-20T00:00:00.000Z", "change": "The SCA limit of removing at most 200 ppm now applies, as in the handbook, to input water above 250 ppm alkalinity (brief, text and decarboniser figure), not to all or 'harder' waters.", "reason": "The handbook ties the 200 ppm limit to alkalinity above 250 ppm; below that it sets 40–70 ppm output." }, { "date": "2026-10-20T00:00:00.000Z", "change": "The SCAA advice on reverse-osmosis systems now says to monitor water pressure to make sure membranes are not clogged; the claim that pressure rises as they clog was removed.", "reason": "The handbook does not say in which direction the pressure changes." } ], "sources": [ { "id": "s1", "title": "The SCA Water Quality Handbook, edition 2 (ISBN 978-0-9995807-3-8)", "authors": "Wellinger, M.; Smrke, S.; Yeretzian, C.", "year": 2018, "publisher": "Specialty Coffee Association", "url": "https://digitalcollection.zhaw.ch/handle/11475/12365", "tier": "A/B", "tierNote": "A for the SCA's own targets and recommendations (recommended ranges, corrosion limits, decarboniser bypass settings, recipe procedures); B for explanations of treatment methods and chemistry", "quote": "Set the bypass valve on your decarbonizer cartridge so that the output water has an alkalinity of 40 -70 ppm CaCO3.", "accessed": "2026-10-20T00:00:00.000Z" }, { "id": "s2", "title": "The Water Quality Handbook (ISBN 978-1-882552-08-5)", "authors": "Beeman, D.; Songer, P.; Lingle, T. (ed. Everage, L.)", "year": 2011, "publisher": "Specialty Coffee Association of America", "tier": "A/B", "tierNote": "A for the SCAA's own standard values; B for treatment advice, trial reports and explanations (all uses in this entry)", "quote": "Since the oxidizing ability of chlorine is removed, bacteria can build up within the filter.", "accessed": "2026-10-20T00:00:00.000Z" }, { "id": "s3", "title": "Hardness of Water", "authors": "U.S. Geological Survey, Water Science School", "publisher": "U.S. Geological Survey", "url": "https://www.usgs.gov/water-science-school/science/hardness-water", "tier": "A", "quote": "When hard water is heated, such as in a home water heater, solid deposits of calcium carbonate can form.", "accessed": "2026-10-20T00:00:00.000Z" }, { "id": "s4", "title": "Hardness in Drinking-water: Background document for development of WHO Guidelines for Drinking-water Quality (WHO/HSE/WSH/09.01/10)", "authors": "World Health Organization", "year": 2009, "publisher": "World Health Organization", "url": "https://hispagua.cedex.es/sites/default/files/hispagua_documento/dureza.pdf", "tier": "A", "quote": "Calcium and magnesium, the two principal ions, are present in many sedimentary rocks, the most common being limestone and chalk.", "accessed": "2026-10-20T00:00:00.000Z" }, { "id": "s5", "title": "Water quality for Espresso coffee", "authors": "Navarini, L.; Rivetti, D.", "year": 2010, "publisher": "Food Chemistry 122(2), 424–428", "url": "https://agris.fao.org/search/en/records/65de275f4c5aef494fda723f", "doi": "10.1016/j.foodchem.2009.04.019", "tier": "A", "quote": "It is well known that water treatment is necessary to remove possible off-flavours deriving from the disinfection performed at municipal waterworks as well as to prevent expensive professional Espresso coffee machine from scaling problems.", "abstractOnly": true, "accessed": "2026-10-20T00:00:00.000Z" }, { "id": "s6", "title": "Process for removing alkalinity and hardness from waters (US Patent 4,235,715)", "authors": "Wiegert, R. E. (Water Refining Company, Inc.)", "year": 1980, "publisher": "United States Patent and Trademark Office", "url": "https://patents.justia.com/patent/4235715", "tier": "A", "tierNote": "A for how the process is built and the reactions it relies on (H⁺-form weak-acid resin taking up hardness matched by alkalinity, carbonic acid breaking down to CO₂ and water, acid regeneration); not used for performance claims", "quote": "The carbonic acid formed by the ion exchange process in the first column breaks down into carbon dioxide and water.", "accessed": "2026-10-20T00:00:00.000Z" }, { "id": "s7", "title": "Sodium (chemical fact sheet), Guidelines for drinking-water quality", "authors": "World Health Organization", "publisher": "World Health Organization", "url": "https://www.who.int/docs/default-source/wash-documents/wash-chemicals/sodium-chemical-fact-sheet.pdf", "tier": "A", "quote": "It should be noted that some water softeners can add significantly to the sodium content of drinking-water.", "accessed": "2026-10-20T00:00:00.000Z" }, { "id": "s8", "title": "Anerkannte Aufbereitungsverfahren für Trinkwasser", "authors": "Bundesamt für Gesundheit (BAG), Abteilung Lebensmittelsicherheit", "year": 2010, "publisher": "Bundesamt für Gesundheit, Bern", "url": "https://www.blv.admin.ch/dam/blv/de/dokumente/lebensmittel-und-ernaehrung/lebensmittelsicherheit/verantwortlichkeiten/anerkannte-aufbereitungsverfahren-trinkwasser.pdf", "tier": "A", "quote": "In der Wasseraufbereitung kann speziell der Ionentausch zu einer signifikanten Erhöhung der Natriumkonzentration führen.", "accessed": "2026-10-20T00:00:00.000Z" }, { "id": "s9", "title": "Effect of point-of-use, activated carbon filters on the bacteriological quality of rural groundwater supplies", "authors": "Snyder, J. W. Jr.; Mains, C. N.; Anderson, R. E.; Bissonnette, G. K.", "year": 1995, "publisher": "Applied and Environmental Microbiology 61(12), 4291–4295", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC167740", "doi": "10.1128/aem.61.12.4291-4295.1995", "tier": "A", "quote": "PAC treatment does not compromise the bacteriological quality of drinking water obtained from underground sources.", "accessed": "2026-10-20T00:00:00.000Z" }, { "id": "s10", "title": "Comparison of the microbiologic quality of point-of-use (POU)-treated water and tap water", "authors": "Chaidez, C.; Gerba, C. P.", "year": 2004, "publisher": "International Journal of Environmental Health Research 14(4), 253–260", "url": "https://experts.arizona.edu/en/publications/comparison-of-the-microbiologic-quality-of-point-of-use-pou-treat/", "doi": "10.1080/09603120410001725595", "tier": "A", "quote": "The highest concentration of bacteria was found in POU-treated water.", "abstractOnly": true, "accessed": "2026-10-20T00:00:00.000Z" }, { "id": "s11", "title": "Safe Drinking-water from Desalination (WHO/HSE/WSH/11.03)", "authors": "World Health Organization", "year": 2011, "publisher": "World Health Organization", "url": "https://fr.ircwash.org/sites/default/files/WHO-2011-Safe.pdf", "tier": "A", "quote": "Desalinated water is low in minerals and is poorly buffered.", "accessed": "2026-10-20T00:00:00.000Z" }, { "id": "s12", "title": "Australian Drinking Water Guidelines, Part 5: Chlorophenols", "authors": "National Health and Medical Research Council (NHMRC)", "publisher": "National Health and Medical Research Council, Australian Government", "url": "https://guidelines.nhmrc.gov.au/australian-drinking-water-guidelines/part-5/physical-chemical-characteristics/chlorophenols", "tier": "A", "quote": "Chlorophenols have taste and odour thresholds in the range 0.0001 mg/L to 0.002 mg/L", "accessed": "2026-10-20T00:00:00.000Z" }, { "id": "s13", "title": "Commission Implementing Decision (EU) 2024/367 laying down rules for the application of Directive (EU) 2020/2184 by establishing the European positive lists of starting substances, compositions and constituents authorised for use in the manufacture of materials or products that come into contact with water intended for human consumption", "authors": "European Commission", "year": 2024, "publisher": "Official Journal of the European Union (copy hosted by FAOLEX)", "url": "https://faolex.fao.org/docs/pdf/eur232271.pdf", "tier": "A", "quote": "are water treatment chemicals and/or filter media and are covered by Article 12 of Directive (EU) 2020/2184", "accessed": "2026-10-20T00:00:00.000Z" }, { "id": "s14", "title": "Compilation of Henry’s law constants (version 5.0.0) for water as solvent – carbon dioxide (CAS RN 124-38-9)", "authors": "Sander, R.", "year": 2023, "publisher": "Atmospheric Chemistry and Physics 23, 10901–12440", "url": "https://henrys-law.org/henry/casrn/124-38-9", "doi": "10.5194/acp-23-10901-2023", "tier": "A", "quote": "3.4×10−4 2300 Burkholder et al. (2019) L", "accessed": "2026-10-20T00:00:00.000Z" }, { "id": "s15", "title": "The solubilities of calcite, aragonite and vaterite in CO2-H2O solutions between 0 and 90°C, and an evaluation of the aqueous model for the system CaCO3-CO2-H2O", "authors": "Plummer, L. N.; Busenberg, E.", "year": 1982, "publisher": "Geochimica et Cosmochimica Acta 46(6), 1011–1040", "url": "https://pubs.usgs.gov/publication/70011789", "doi": "10.1016/0016-7037(82)90056-4", "tier": "A", "quote": "At 25°C the logarithms of the equilibrium constants are −8.480 ± 0.020, −8.336 ± 0.020 and −7.913 ± 0.020", "abstractOnly": true, "accessed": "2026-10-20T00:00:00.000Z" }, { "id": "s16", "title": "Water for Extraction – Composition, Recommendations, and Treatment", "authors": "Wellinger, M.; Smrke, S.; Yeretzian, C.", "year": 2017, "publisher": "In: Folmer, B. (ed.), The Craft and Science of Coffee, Academic Press, ch. 16 (pp. 391–396 cited)", "tier": "B", "quote": "Other methods that can even further purify water are distillation and precipitation.", "accessed": "2026-10-20T00:00:00.000Z" }, { "id": "s17", "title": "Wasserhärte: Was muss beachtet werden? (SVGW Trinkwasserinformation TWI 13/2008)", "authors": "SVGW (Schweizerischer Verein des Gas- und Wasserfaches)", "year": 2008, "publisher": "SVGW, Information Trinkwasser (copy hosted by Kanton Solothurn)", "url": "https://so.ch/fileadmin/internet/ddi/ddi-gesa-lk/Umgang_mit_hartem_Wasser__Wasserhaerte__Entkalkungsanlagen_.pdf", "tier": "B", "quote": "Wird eine Enthärtung vorgenommen, soll eine Resthärte von 12–15 ºfH nicht unterschritten werden.", "accessed": "2026-10-20T00:00:00.000Z" } ], "markdown": "## Why water is treated\n\nTap water is made safe to drink, not to brew coffee. Treatment for coffee has three aims: to remove off-flavours, to protect machines and to adjust the minerals for flavour.1,5\n\n**Off-flavours.** Waterworks disinfect water with chlorine or chloramines, and both leave a strongly unpleasant flavour in the brew.1 Chlorine can also react with phenols from the coffee to form chlorophenols, which taste medicinal.2 In drinking water, chlorophenols are noticed at extremely low levels.12 Iron, earthy and mouldy notes are other taints that treatment may have to remove.1\n\n**Scale.** When hard water is heated, calcium carbonate settles out as scale.3 Scale slows heat transfer and clogs valves and jets, especially where water is hottest.1 A review of espresso water calls treatment necessary for both reasons, taints and scale.5\n\n**Corrosion.** Water that is too soft or poorly buffered attacks metal instead.1,4 The SCA links corrosion to alkalinity below 40 ppm as CaCO₃, pH below 6 or above 8, and much chloride, sulfate or nitrate.1 Treatment must therefore not go too far. The water entry sets out the corrosion rules in detail.\n\n**Flavour.** Alkalinity, the water’s capacity to neutralise acid, decides most of how water changes the cup; hardness matters less.1 According to the SCA handbook, most tap waters in central Europe and the US carry more hardness and alkalinity than suits coffee.1 Hardness and alkalinity, and how they shape flavour, have their own entries.\n\n## What each treatment does\n\nA treatment either removes something, swaps one ion for another, or adds minerals. Each one moves the water in its own direction on a chart of alkalinity against total hardness.1\n\n| Method | How it works | Total hardness | Alkalinity | Also changes | Source |\n|---|---|---|---|---|---|\n| Particle (sediment) filter | Traps sand, rust and other particles; filters can run in series, coarse to fine | – | – | Very fine filters clog sooner | SCAA2 |\n| Activated carbon | Adsorbs chlorine, chloramines and organic matter | – | – | Removes the chlorine that kept bacteria in check | SCA1, SCAA2 |\n| Softener | Swaps Ca²⁺ and Mg²⁺ for Na⁺ or K⁺ | Down | Unchanged | Adds sodium or potassium | SCA1, WHO7 |\n| Decarboniser (weak-acid cation exchanger) | Swaps Ca²⁺ and Mg²⁺ for H⁺; H⁺ turns bicarbonate into carbonic acid | Down | Down by the same amount | Dissolved CO₂ up, pH down | SCA1, patent6 |\n| Buffered decarboniser | Mostly decarboniser resin with a little softener resin | Down | Down, but not to zero | As above, less strongly | SCA1 |\n| Reverse osmosis | A membrane passes water and holds back almost all dissolved substances | Towards zero | Towards zero | Waste water; aggressive if not remineralised | SCA1, SCAA2 |\n| Deioniser | Cation plus anion exchanger removes all ions | Towards zero | Towards zero | Low capacity | SCA1 |\n| Dealkaliser | Anion exchanger or a strong acid | Unchanged | Down | Chloride up (hydrochloric acid or chloride-form resin) | SCA1, textbook16 |\n| Hardener | Adds calcium or magnesium as sulfate or chloride | Up | Unchanged | Sulfate or chloride up | SCA1 |\n| Alkaliser | Dissolves sodium or potassium hydrogen carbonate | Unchanged | Up | Sodium or potassium up | SCA1 |\n| Carboniser | Dissolves magnesium carbonate | Up | Up by the same amount | – | SCA1 |\n| Calcium-for-magnesium exchanger | Swaps one hardness ion for the other | Unchanged | Unchanged | Ca:Mg ratio | SCA1 |\n\nA coffee science textbook chapter adds distillation and precipitation as further methods, but calls them expensive or elaborate and not common for coffee water.16 Mixing two waters gives any composition on the straight line between them.1\n\nThe chart below starts from one of the SCA’s worked examples, a tap water of 150 ppm alkalinity and 200 ppm total hardness.1 Only two of the four routes shown reach the SCA range.\n\n
\n
\n
Four treatments applied to one tap water (150 ppm alkalinity, 200 ppm total hardness), against the SCA range and its corrosion limit. The key shows the direction of every treatment.1,6 End points are our own arithmetic from SCA worked examples. Schematic directions.
\n
\n\n- **Softener** (1): hardness falls, but alkalinity stays at 150 ppm, twice the SCA’s upper limit of 75 ppm.1\n- **Decarboniser with bypass** (2): with 40% of the water bypassed, the mix has 60 ppm alkalinity and 110 ppm hardness (our calculation).\n- **Blending with RO water** (3): 62.1% of almost pure water (5|10 ppm) gives 60 ppm alkalinity and 82 ppm hardness. This is our calculation with the SCA’s mixing formula.1\n- **Dealkaliser** (4): alkalinity falls to 60 ppm, but hardness stays at 200 ppm, above the SCA’s 175 ppm.1\n\n## Choosing a treatment\n\nThe starting water decides the method. The SCA handbook gives these rules:1\n\n- **Limestone water** – alkalinity equal to hardness or up to about 70 ppm below it: a decarboniser suits most uses.\n- **Gypsum water** (alkalinity more than about 70 ppm below hardness) and **salt water**: reverse osmosis is the only reliable method. The same applies to water with iron, manganese or much sodium.\n- **Small volumes at home**: dilute tap water with almost pure water, such as reverse-osmosis water or water below 20 ppm hardness and alkalinity.\n- **No single method reaches the target**: combine a decarboniser with a hardener, or demineralise fully and then add minerals back.\n\nEvery water that passes a decarboniser or dealkaliser should keep at least 40 ppm alkalinity, to avoid corrosion.1 Espresso tolerates more alkalinity than filter coffee, up to about 150 ppm, because it uses far less water per gram of coffee.1 The water entry explains why.\n\n**Espresso machines and kettles differ.** In a closed line to an espresso machine, the carbon dioxide formed by a decarboniser cannot escape.1,16 The SCA therefore advises setting the bypass so that output alkalinity is 40–70 ppm for input waters up to 250 ppm alkalinity. Above 250 ppm, alkalinity should fall by no more than 200 ppm.1 In an open kettle the carbon dioxide escapes on heating, and decarbonisation can be used without these limits.1\n\n**Taints come first.** Activated carbon removes chlorine compounds and organic matter efficiently.1 Low levels of iron can be removed by a softener or decarboniser; higher levels need aeration and filtration of the iron particles that form.1 Of the methods in the SCA handbook, only reverse osmosis and deionisation fully remove iron and lead.1\n\n**In-line or batch.** In-line systems connect the tap to the machine and run without attention. Batch systems treat water in a tank; an espresso machine may then need an extra pump.1 As of 2018, dosing acid or sodium hydrogen carbonate was not sold as an in-line cartridge.1\n\n## Keeping treatment working\n\nEvery cartridge and membrane has a limited capacity. The SCA notes that makers of treatment products often publish tables of capacity by starting water and volume.1 An estimate of monthly or yearly water use then tells when the next change is due.1\n\n**Checking the output.** Tap water can vary by day and season, especially from several sources, shallow groundwater, rivers or lakes.1 A conductivity meter can track a decarboniser or an RO system with bypass, because their output conductivity changes with treatment.1 A softener barely changes conductivity, so the meter cannot track it reliably.1 The measuring entry covers the methods. For RO, the SCAA handbook recommends monitoring the water pressure to make sure the membranes are not clogged.2\n\n**Hygiene.** Activated carbon removes the chlorine that kept bacteria in check, so bacteria can build up inside the filter.2 A 2004 study found more bacteria in water from household tap filters than in unfiltered tap water.10 A 1995 study found raised bacterial counts after water stood overnight in carbon filters, but lower counts than the incoming water after a short flush.9 Bacteria have also been found in reverse-osmosis permeate, and they can multiply in the lines downstream.11 According to the SCAA handbook, carbon filters must be maintained or replaced regularly.2 After a boil-water alert, the handbook advises sanitising the lines behind the filter, because the filter removes the extra chlorine the waterworks adds.2\n\n\n\n## Ion exchange chemistry\n\nIon exchangers are resin beads that hold one kind of ion and release another. Softeners and decarbonisers both take up calcium and magnesium; they differ in what they give back.1\n\n**Softener.** The resin releases two sodium (or potassium) ions for each calcium or magnesium ion it takes up:1\n\n2 R–Na + Ca²⁺ → R₂–Ca + 2 Na⁺\n\nBicarbonate is untouched, so alkalinity does not change.1 Water from limestone keeps its full buffer, now paired with sodium. Sodium rises in step: one Swiss federal guidance names ion exchange specifically as a cause of significantly higher sodium,8 and the World Health Organization (WHO) notes that some softeners add significantly to it.7 Our own calculation: each 1 ppm of hardness removed (as CaCO₃) adds 0.46 mg/L of sodium, so softening 200 ppm adds about 92 mg/L. The WHO sets no health-based value for sodium but notes that more than 200 mg/L may give an unacceptable taste.7\n\n**Decarboniser.** A weak-acid cation resin in hydrogen form releases hydrogen ions instead:1,6\n\n2 R–H + Ca²⁺ → R₂–Ca + 2 H⁺ \nH⁺ + HCO₃⁻ → H₂CO₃ ⇌ CO₂ + H₂O\n\nThe resin removes hardness only to the extent that it is matched by alkalinity; in one patented process a second, sodium-form column then softens the rest.6 Hardness and alkalinity therefore fall by equal amounts.1 The carbonic acid breaks down into carbon dioxide and water.6 In the patented process, exhausted resin is regenerated with a mineral acid, preferably hydrochloric acid.6 A buffered decarboniser mixes in a small share of softener resin, so that some alkalinity remains.1\n\n
\n
\n
A decarboniser with bypass: the resin swaps calcium and magnesium for hydrogen ions, which turn bicarbonate into carbonic acid; the bypass sets how much alkalinity remains.1,6 The mixing rule is our own calculation. Schematic, not to scale.
\n
\n\n**The bypass as a mass balance.** If the resin removes all the alkalinity it treats, output alkalinity equals the bypassed share times input alkalinity (our calculation). Hardness that is not matched by alkalinity passes the resin unchanged.6 For the SCA example of 150 ppm alkalinity and 200 ppm hardness, a 40% bypass gives 0.4 × 150 = 60 ppm alkalinity. Total hardness becomes 0.4 × 200 + 0.6 × 50 = 110 ppm (our calculation). Real cartridges, especially buffered ones, leave some alkalinity in the treated stream, so the setting differs in practice.1\n\n**Carbon dioxide from decarbonisation.** Each hydrogen carbonate ion neutralised becomes one molecule of carbonic acid, which is in exchange with dissolved CO₂.1 Our own calculation: 1 ppm of alkalinity removed (as CaCO₃) is 0.020 mmol/L of hydrogen carbonate and forms 0.88 mg/L of CO₂. Removing 90 ppm, as in the example above, forms about 79 mg/L; removing 200 ppm forms 176 mg/L; removing 250 ppm forms 220 mg/L. In a closed line this CO₂ lowers the pH and raises the corrosion risk.1 According to the SCA handbook and a textbook chapter, it also adds to the CO₂ from fresh coffee and can give espresso a foamy crema with large, short-lived bubbles.1,16 The chapter names degassing, or a demineraliser that forms no extra CO₂, as remedies.16\n\n**Dealkalisers.** An anion exchanger can swap hydrogen carbonate for chloride; the textbook chapter described it in 2017 as not yet commercially available for coffee.16 Adding a strong acid works the same way on alkalinity.1,16 Both raise chloride, which the SCA links to corrosion in espresso machines.1\n\n## Membranes, blending and remineralisation\n\n**Reverse osmosis.** Pressure drives water through a semi-permeable membrane; dissolved minerals pass far less readily and are flushed to waste with part of the water.1,2 According to the SCAA handbook, this wastes considerable water and makes RO the most expensive treatment it considers.2 A deioniser reaches a similar result by ion exchange, but the SCA finds its capacity too low for high-volume commercial use.1\n\n**Why pure water is not the end point.** Desalinated water is low in minerals, poorly buffered and usually aggressive to metals and cement.11 Water supplies therefore remineralise it, often with calcium carbonate, to reduce its corrosive potential.11 The SCAA handbook warns that RO water can corrode metal pipes unless the system is engineered and installed properly, and rules it out with galvanised pipes or lead-soldered copper.2 It adds that water with almost no minerals may not be detected by the level probes in many brewers.2\n\n**Blending.** Mixing RO water with tap water gives any composition on the line between the tap water and zero.1 The share of water 1 for a target alkalinity is (alkalinity of water 2 − target) ÷ (alkalinity of water 2 − alkalinity of water 1); the same formula works for hardness.1 The water entry corrects a slip in the handbook’s worked example.\n\n**Using the concentrate.** Some RO systems can deliver the concentrate instead of the permeate, raising hardness and alkalinity without changing their ratio.1 Our own calculation, assuming the membrane holds back all minerals: the concentration factor is the feed volume divided by the concentrate volume, so a system that sends half the feed to concentrate doubles the minerals. The SCA restricts this use to low-mineral source water that is otherwise of high quality, free of metals and chlorine.1\n\n## Making water from salts\n\nRecipe waters start from almost pure water and add salts. The SCA handbook gives stock concentrates; we checked each with molar masses.1\n\n| Concentrate | Salt per 1 L of RO water | Strength stated | Dose stated | Our check |\n|---|---|---|---|---|\n| Alkalising | 3 g sodium hydrogen carbonate | 100 °dH alkalinity | 10 g per L adds 1 °dH | 35.7 mmol/L HCO₃⁻ = 100 °dH |\n| Magnesium hardening | 7.3 g MgCl₂·6H₂O | 200 °dH hardness | 10 g per L adds 2 °dH | 35.9 mmol/L Mg²⁺ ≈ 201 °dH |\n| Calcium hardening | 4.0 g CaCl₂ | 200 °dH hardness | not stated | 36.0 mmol/L Ca²⁺ ≈ 202 °dH (anhydrous salt) |\n| Dealkalising (not advised without training) | 3 g hydrochloric acid, 37% | −100 °dH alkalinity | 10 g per L removes 1 °dH | 30.4 mmol/L H⁺ ≈ −85 °dH |\n\nSources: SCA1 for the recipes; checks are our own calculation (1 °dH = 0.178 mmol/L Ca²⁺ or Mg²⁺, or 0.357 mmol/L HCO₃⁻; molar masses NaHCO₃ 84.01, MgCl₂·6H₂O 203.30, CaCl₂ 110.98, HCl 36.46 g/mol). Three of the four recipes agree with their stated strength. By our arithmetic the acid recipe is about −85 °dH rather than −100 °dH, so 10 g of it removes about 0.85 °dH per litre, not 1 °dH. The SCA advises against handling concentrated acid without training and protective equipment.1\n\n**Sulfate or chloride.** For espresso machines the SCA advises calcium or magnesium sulfate rather than chlorides, because chlorides increase the chance of corrosion.1 Baking soda sold for cooking often contains other ingredients, so the SCA asks for pure sodium hydrogen carbonate.1\n\n**Three waters.** Mixing three waters lets hardness and alkalinity be set independently.1 In the SCA example, an almost pure water (10|10), a hard mineral water (300|320) and an Epsom salt concentrate (10|450) are mixed.1 For a target of 40 ppm alkalinity and 55 ppm hardness, the handbook gives 867.2 g, 103.4 g and 29.4 g per litre. For 60|80 it gives 790.0 g, 172.4 g and 37.6 g.1 Our check of the second: 0.790 × 10 + 0.172 × 300 + 0.038 × 10 = 60 ppm alkalinity, and 0.790 × 10 + 0.172 × 320 + 0.038 × 450 = 80 ppm hardness.\n\nHow individual minerals change flavour is covered in the entry on water minerals and flavour.\n\n## Scale chemistry and scale inhibitors\n\n**Why heat makes scale.** Calcite, the common form of calcium carbonate, dissolves only slightly; its solubility product at 25 °C has a logarithm of −8.480.15 Tap water holds calcium and hydrogen carbonate in solution because dissolved CO₂ keeps the equilibrium on the dissolved side.1 Heating drives CO₂ out, because gases dissolve less in hot water.1,14 Our own rough estimate from Henry’s law (constant 3.4 × 10⁻⁴ mol/(m³·Pa) at 25 °C, temperature coefficient 2300 K):14 at 95 °C, CO₂ is only about one fifth as soluble as at 20 °C. As CO₂ leaves, calcium carbonate precipitates.1,3 At most the lower of hardness and alkalinity can leave as scale, which is why both treatments that lower either one reduce scale.1\n\n**Scale inhibitors.** Polyphosphates such as hexametaphosphate bind a limited amount of hardness and keep it from forming scale.2 According to the SCAA handbook, they help little in hard water, work less well at high pH and do nothing against silica.2 They do not lower hardness or alkalinity, so they do not change the water for flavour (our inference from the mechanism described).\n\n**Magnetic devices.** According to the SCAA handbook, a review by a water-treatment trade association found too little study to certify claims for magnetic scale prevention. The handbook argues that any altered crystal form would not survive the rapid heating in a brewer.2\n\n## Rules for treatment devices\n\nTreatment devices touch drinking water, so drinking-water law applies to them. These are rules, cited as rules; they say nothing about coffee quality.\n\n- **European Union.** Directive (EU) 2020/2184 sets requirements for materials in contact with drinking water and, separately, for water treatment chemicals and filter media. A 2024 implementing decision states that membranes and ion-exchange resins fall under the article on treatment chemicals and filter media.13\n- **Switzerland.** A 2010 guidance from the Federal Office of Public Health lists recognised treatment processes for drinking water and notes that ion exchange can raise sodium significantly.8 The SVGW, the Swiss water industry association, recommends keeping a minimum residual hardness when drinking water is softened.17\n- **Taste limits.** Australian drinking-water guidelines set aesthetic limits for chlorophenols at their taste thresholds, which lie below their health-based values.12\n\nNo European standard for softeners or point-of-use filters was opened for this entry.\n\n## Reference values\n\n| Parameter | Value reported | Source |\n|---|---|---|\n| Minimum alkalinity after decarbonisation or dealkalisation | 40 ppm as CaCO₃ | SCA 20181 |\n| Decarboniser output, espresso, input ≤ 250 ppm alkalinity | 40–70 ppm alkalinity | SCA 20181 |\n| Decarboniser, input > 250 ppm alkalinity | Remove ≤ 200 ppm | SCA 20181 |\n| Decarboniser choice | Alkalinity equal to hardness or up to ≈70 ppm below it | SCA 20181 |\n| Alkalinity tolerated in espresso | Up to ≈150 ppm | SCA 20181 |\n| CO₂ formed per ppm alkalinity removed | 0.88 mg/L (our calculation) | from the reaction1,6 |\n| Sodium added per ppm hardness softened | 0.46 mg/L (our calculation) | from ion exchange1 |\n| Sodium, taste | > 200 mg/L may give unacceptable taste | WHO7 |\n| Chlorophenols, taste and odour thresholds in water | 0.0001–0.002 mg/L | NHMRC12 |\n| Calcite solubility product, 25 °C | log K = −8.480 ± 0.020 | Plummer & Busenberg15 |\n| CO₂ solubility, 95 °C relative to 20 °C | ≈0.2 (our estimate) | from Henry’s law14 |\n| RO concentrate, minerals | × feed ÷ concentrate volume (our calculation; full rejection assumed) | SCA principle1 |\n\n## Open questions\n\n- **Calcium-for-magnesium exchange.** Cartridges that swap calcium for magnesium change neither hardness nor alkalinity. Calcium and magnesium may affect extraction differently, but the SCA handbook calls this an area of recent research.1 No sensory study of such cartridges was found for this entry.\n- **Hygiene of café cartridges.** The microbiology studies opened here tested household devices on drinking water.9,10 No study opened for this entry measured bacteria in coffee-machine filter cartridges.\n- **Decarbonised water and machine life.** The SCA limits rest on corrosion indices and calculation.1 No long-term field study of corrosion in espresso machines run on decarbonised water was found.\n- **Salt choice and flavour.** Sulfate is advised over chloride for machine safety.1 Whether the anion changes cup flavour at recipe-water levels was not tested in any study opened for this entry.\n\n## Key studies\n\n
\n

Snyder et al., 1995 – Carbon filters on rural groundwater. Monitored point-of-use powdered activated carbon filters on 24 rural groundwater supplies, sampling raw, filtered and overnight-standing water. Heterotrophic bacteria rose after stagnation in the filters but fell below incoming levels after a brief flush, and coliforms were reduced; the authors judged that, apart from periods after stagnation, the filters did not compromise water quality.

\n

Applied and Environmental Microbiology 61(12), 4291–4295 · doi:10.1128/aem.61.12.4291-4295.1995

\n
\n\n
\n

Chaidez & Gerba, 2004 – Tap filters and bacteria. Compared water from household carbon filters with tap water from the same and from unfiltered taps. Filtered water carried the highest bacterial counts, which the authors suggest may come from biofilm growth in the devices; they estimated a low risk of colonisation by the opportunistic bacteria found.

\n

International Journal of Environmental Health Research 14(4), 253–260 · doi:10.1080/09603120410001725595

\n
\n\n
\n

Navarini & Rivetti, 2010 – Water quality for espresso. Reviews how water and coffee components interact in espresso brewing. It starts from the two accepted reasons for treatment, disinfection taints and machine scale, and argues that the direct effect of water composition on the cup deserves more attention.

\n

Food Chemistry 122(2), 424–428 · doi:10.1016/j.foodchem.2009.04.019

\n
\n\n
\n

Plummer & Busenberg, 1982 – Calcium carbonate solubility. Measured the solubility of calcite, aragonite and vaterite in water with carbon dioxide from 0 to 90 °C, in about 350 new measurements. It gives equilibrium constants for all three forms and a revised model of the dissolved species.

\n

Geochimica et Cosmochimica Acta 46(6), 1011–1040 · doi:10.1016/0016-7037(82)90056-4

\n
" }