CoffeariumBeta
  1. Plant
  2. Farm
  3. Process
  4. Trade
  5. Roast
  6. Brew
  7. Cup
Equipment · Espresso

Espresso machine care

What builds up inside an espresso machine – scale, milk residues and bacteria – what tests found about metals released after descaling, and what the evidence does and does not say about cleaning.

Across the chain: Equipment

Source-checkedby Coffearium verification pass· Last source-checked 27 Oct 2026· Prepared with AI assistance· 13 sources (8 tier A)· 1 contested· Updated 27 Oct 2026· Literature last searched 27 Oct 2026
Reading depth
Contents
  • Scale forms when heat drives calcium carbonate out of hard water; it lowers heat transfer and clogs valves and jets, especially in the hot-water sections.1,2
  • In tests of eight new coffee machines, portafilter machines released the most metals, and release rose sharply after descaling.5
  • The German Federal Institute for Risk Assessment (BfR) recommends rinsing machines thoroughly after descaling and, in daily use, a water-only rinse before preparing espresso or coffee.4
  • In 2023 and 2024, two Swiss cantonal laboratories found milk from fully automatic coffee machines above hygiene guideline values in 4 of 19 and at least 4 of 15 samples.9,10
  • No study consulted for this entry tested backflushing, group cleaning detergents or the replacement of gaskets and shower screens.

What builds up inside the machine#

An espresso machine keeps water hot, often for hours, and passes it through metal parts, coffee and milk.1,5 Several things build up or wear as a result. Hard water leaves scale, a mineral crust, in boilers and narrow passages.1,2 Water that is too soft or acidic attacks metal instead, and metal can then pass into the water.1,3 Milk leaves residues in steam wands and milk systems.10,13 Bacteria can grow wherever residues and water stay together.9,10,11

How the machine itself works is covered in How an espresso machine works. Water composition, hardness and treatment have their own entries.

13247568
  1. Water filter. Treats the supply; carbon filters can host bacteria once the chlorine is gone.3
  2. Boiler and heating element. Where heat makes scale; a scale layer insulates the heater.1,2
  3. Flow restrictor (gicleur). A narrow jet in the water line; scale can block it.2
  4. Three-way valve and drain line. Opens the group to a drain at atmospheric pressure at the end of the shot.12
  5. Group and shower screen. The shower head spreads water over the coffee.12
  6. Portafilter and basket. Portafilter machines released the most metals in one study.5
  7. Steam wand. Heats and froths milk; milk residues stick to it.13
  8. Drip tray. Waste coffee in capsule machines was colonised by bacteria within days.11
A single-boiler espresso machine cut along its centre line, with the parts discussed in this entry. The accent marks scale in the boiler, the narrow restrictor and the wand tip. Schematic, not to scale.

Scale and descaling#

Why scale forms. Tap water carries dissolved calcium and hydrogen carbonate. Heat drives off carbon dioxide, and calcium carbonate settles out as scale.1 In steam boilers, at a pH above 10, magnesium hydroxide can form as well, especially with sodium-softened water.1,2 Silica scales too, whatever the pH and alkalinity.3 The chemistry is explained in Hardness and alkalinity.

What scale does. A layer of scale insulates the heating element, so heating becomes less efficient.2 Scale also clogs valves, jets and flow restrictors, especially in the hot-water sections.1,2 According to a water handbook of the former Specialty Coffee Association of America (SCAA), scale grows back faster once a first deposit has formed. Even after descaling, the rough surface left behind speeds up new scale, and the handbook suggests replacing the part.3

Descaling. Descaling dissolves scale with an acid. A coffee science textbook chapter states that machine companies recommend regular descaling with an acid dissolved in water, preferably one with little taste.2 Prevention starts with the water: lower alkalinity and a lower pH mean less scale.1 Water treatment covers filters and softeners.

Metals released from machines#

Metal parts of a machine meet hot water, coffee and descaling acid.4,5 In a 2015 German laboratory study, 8 coffee machines and 11 electric kettles were filled with food simulants, and the released metals were measured.5 Portafilter machines released by far the most, in amount and in the number of elements.5 Release fell during each day and over consecutive days. After descaling it rose sharply.5

The BfR considers it possible that acidic descalers release lead from metal components.4 It recommends rinsing the machine thoroughly after descaling. For daily use, it also recommends a rinse with water only before each preparation of espresso or coffee, as manufacturers also prescribe.4 The BfR states that its sample was small and not representative.4 Food-safety limits for metals are set out in Contaminants and food safety.

In the cup. A 2025 survey sampled espresso from cafés, restaurants and petrol stations in Poland and Portugal.8 Nickel and lead were higher in the espresso than in the tap water at the same places.8 The authors could not say whether the nickel came from the coffee or from the machine.8

Coffee and milk residues#

Group, shower screen and valve. In the group, the shower head spreads water over the coffee bed.12 According to one patent, the valve that ends the shot is typically a three-way valve. It opens the group to a drain line at atmospheric pressure, so the leftover pressure escapes.12 Cleaning compounds that are not rinsed off can dissolve into the next brew, according to the SCAA handbook.3

No study consulted for this entry tested backflushing, group detergents or how often gaskets and shower screens should be replaced. These routines are set by machine makers; their instructions were not consulted. How coffee oil goes rancid is covered in Coffee oil.

Steam wands and milk systems. A steam wand heats and froths milk with steam from the boiler, and milk residues tend to stick to it.13 Swiss food inspectors tested milk and milk drinks from fully automatic coffee machines with milk functions. In Basel-Stadt in 2023, 4 of 19 samples were above guideline values.9 In Zürich in 2024, 4 of 15 samples were above the guideline value for Enterobacteriaceae.10 Both laboratories named poorly cleaned milk-carrying parts as one cause.9,10

Drip trays and waste. In a 2015 study of capsule machines, bacteria colonised the waste coffee within days.11 The community settled into a more stable mix after about a month.11

Water filters#

A filter changes what reaches the boiler, so it changes how much scale forms. The options and their limits are covered in Water treatment. Filters need care of their own. Activated carbon removes the chlorine that kept bacteria in check, so bacteria can build up inside the filter.3 Treatment must also not go too far: the SCA links corrosion to alkalinity below 40 ppm as CaCO₃.1

Scale in brew and steam boilers#

The Specialty Coffee Association (SCA) water handbook calculates the most scale that can form at 95 °C and at 130 °C, which it pairs with 1.7 bar of steam pressure.1 It uses a modified Langelier index with an equilibrium pH based on alkalinity.1 Its core zone for espresso machines allows at most 12 mg/L of scale in a brew boiler and 30 mg/L in a steam boiler at 130 °C.1 How these values turn into grams of scale per year is worked through in Water for coffee.

The calculation gives a maximum. At most the lower of total hardness and alkalinity can leave the water as calcium carbonate.1 According to the SCA handbook, lower pH means less scale.1 Magnesium hydroxide scale needs a pH above 10.1 A textbook chapter places it in steam boilers, especially with sodium-softened water.2

Corrosion and metal release#

Corrosion from water. The SCA links corrosion to alkalinity below 40 ppm as CaCO₃, to a pH below 6 or above 8 and to high chloride, sulfate or nitrate.1 The SCAA handbook reports corroded stainless-steel brewers in which deposits held high chloride, iron and nickel.3 It traced the usual cause to acid flux from soldering copper pipe joints, not to the water.3 The same handbook notes that brass pumps often corrode with unsuitable water.3

Release into food simulants. In the 2015 German study, releases from portafilter machines of lead, nickel, manganese, chromium and zinc lay within and partly beyond the Council of Europe release limits.5 One portafilter machine released much less than the others; the authors take this as evidence that release above the level of health concern is technically avoidable.5 The kettles in the same study released little.5 The authors recommend a careful rinsing routine, especially after descaling.5 The study used food simulants without coffee, so the release it measured came from the machines. The full text was not consulted; the values per machine are not given here.

Earlier surveys. A Danish study of 2000 found that coffee machines released no aluminium, lead, chromium or nickel in quantities of any significance.7 The BfR reports two surveys by the state laboratory in Stuttgart. In 2007, 3 of 17 machines showed increased lead release.4 In 2011, no lead release was detected, with a limit of determination of 0.005 mg/L of test water.4

Legal position. As of 2015 and 2025, the EU had no legal limit for metals released from metallic food contact materials.4,6 In 2013 the Council of Europe recommended release limits for 21 metals and metalloids (Resolution CM/Res(2013)9).6 For lead, it sets a specific release limit of 10 µg per kg of food, a recommendation rather than law.4

Metals in café espresso. The 2025 survey measured 60 espresso samples in Poland and 44 in Portugal by inductively coupled plasma mass spectrometry (ICP-MS), with tap water from the same places.8 Median nickel was 81.0 µg/L in Polish and 86.8 µg/L in Portuguese espresso, against 5.58 and 4.61 µg/L in the water.8 Median lead was 14.3 and 4.43 µg/L in espresso, against 2.15 and 0.574 µg/L in the water.8 The authors point to other studies in which nickel and lead leached from machines, especially portafilter machines after descaling.8 For nickel, they state that their method could not identify whether it came from the coffee powder or the machine; for lead, they discuss coffee composition, extraction conditions and machine build and maintenance as possible reasons for the differences.8 They add that copper does not seem to leach from coffee machines in significant amounts.8

020406080100120140160µg/LBars: 25th–75th percentile of the samples.Nickel in espresso, Poland(median 81.0)45.3–1338 Nickel in espresso,Portugal(median 86.8)58.0–1218 Nickel in tap water,Poland(median 5.58)4.08–9.628 Nickel in tap water,Portugal(median 4.61)1.09–6.558 Lead in espresso, Poland(median 14.3)2.49–34.08 Lead in espresso, Portugal(median 4.43)1.99–15.68 Lead in tap water, Poland(median 2.15)1.58–3.898 Lead in tap water,Portugal(median 0.574)0.254–1.378
Data table
Pointµg/LSource
Nickel in espresso, Poland (median 81.0)45.3–13345.3–1338
Nickel in espresso, Portugal (median 86.8)58.0–12158–1218
Nickel in tap water, Poland (median 5.58)4.08–9.624.08–9.628
Nickel in tap water, Portugal (median 4.61)1.09–6.551.09–6.558
Lead in espresso, Poland (median 14.3)2.49–34.02.49–348
Lead in espresso, Portugal (median 4.43)1.99–15.61.99–15.68
Lead in tap water, Poland (median 2.15)1.58–3.891.58–3.898
Lead in tap water, Portugal (median 0.574)0.254–1.370.254–1.378
Nickel and lead in espresso served in Poland (60 samples) and Portugal (44 samples), and in tap water from the same places; bars show the middle half of the samples, medians in the labels. The study did not identify whether the metals came from the coffee or the machine.8

Descaling agents#

The sources consulted describe descaling in general terms only. A textbook chapter states that machine companies recommend an acid dissolved in water, preferably one with little taste, and sell their own solutions.2 The BfR considers it possible that acidic descalers release lead from metal parts.4 In the 2015 study, release rose sharply after descaling.5 The BfR and the study authors both recommend thorough rinsing afterwards.4,5

No study consulted for this entry compared citric, lactic and sulfamic acid descalers for scale removal, attack on metal parts or residues in the cup. The descaling routine of each machine is set by its maker; those instructions were not consulted for this entry.

Milk systems and hygiene#

The two Swiss surveys judged milk against guideline values of the Swiss hospitality guideline, not against legal maximum values.9,10 In Basel-Stadt, 4 of 19 samples (21%) exceeded guideline values in one or two parameters.9 The guideline value for Enterobacteriaceae is 100 colony-forming units (CFU) per gram, and Bacillus cereus counts stayed below 5,000 CFU/g in both samples that exceeded its guideline value.9 The laboratory named poor cleaning of milk-carrying parts, poor cooling and long or unhygienic storage of milk as possible causes.9 In Zürich, 4 of 15 samples exceeded the guideline value for Enterobacteriaceae, and in one further case the aerobic mesophilic count was above its guideline value.10 Enterobacteriaceae in heated drinks pointed to contamination after heating, for example inside the machine.10 Follow-up checks found no further exceedances, on a spot-check basis.10

Both surveys concern machines with built-in milk functions, not steam wands operated by a barista. No study consulted for this entry measured bacteria on steam wands. A patent for a wand-cleaning device states as background that milk residues stick to wands and serve as a platform for bacterial growth.13

In the capsule-machine study, enterobacteria of the genus Pantoea peaked 4–11 days after a new machine was first used, before a more balanced community formed after one month.11

Reference values#

Parameter Value reported Context Source
Scale calculation temperatures 95 °C; 130 °C at 1.7 bar Brew water; steam boiler SCA 20181
Maximum scale, core zone 12 mg/L (brew boiler); 30 mg/L (steam boiler, 130 °C) SCA recommendation SCA 20181
Corrosion risk Alkalinity below 40 ppm as CaCO₃; pH below 6 or above 8 SCA thresholds SCA 20181
Magnesium hydroxide scale pH above 10 Steam boilers, softened water SCA 20181, Wellinger et al.2
Lead release limit 10 µg/kg food Council of Europe recommendation; no legal EU limit BfR4,6
Machines with increased lead release 3 of 17 (2007); none detected, limit 0.005 mg/L (2011) State laboratory Stuttgart, as reported BfR4
Nickel in espresso, median (25th–75th percentile) 81.0 (45.3–133) µg/L; 86.8 (58.0–121) µg/L Poland; Portugal, cafés Konieczyński et al.8
Lead in espresso, median (25th–75th percentile) 14.3 (2.49–34.0) µg/L; 4.43 (1.99–15.6) µg/L Poland; Portugal, cafés Konieczyński et al.8
Nickel in tap water, median (25th–75th percentile) 5.58 (4.08–9.62) µg/L; 4.61 (1.09–6.55) µg/L Same places Konieczyński et al.8
Lead in tap water, median (25th–75th percentile) 2.15 (1.58–3.89) µg/L; 0.574 (0.254–1.37) µg/L Same places Konieczyński et al.8
Milk samples above guideline values 4 of 19 (21%) Basel-Stadt, 2023 Kantonales Laboratorium Basel-Stadt9
Milk samples above the Enterobacteriaceae guideline 4 of 15 (one further sample above the aerobic mesophilic guideline) Zürich, 2024 Kantonales Labor Zürich10
Enterobacteriaceae guideline value 100 CFU/g Swiss hospitality guideline, as applied Kantonales Laboratorium Basel-Stadt9

Open questions#

  • Descaling agents compared. No study consulted for this entry compared acids for scale removal, metal release or residues, or tested how well descalers remove silica scale.
  • Where the metal comes from. The BfR did not establish the source of the lead and assumes that internal components or soldered joints may contribute4; the café survey could not identify whether the nickel came from the coffee or the machine.8
  • Commercial machines. The BfR tests used new household machines.4 No study consulted for this entry measured release from café machines after descaling.
  • Group cleaning. No study consulted for this entry measured coffee residues in groups, the effect of backflushing on flavour or detergent residues in espresso.
  • Steam wands. The microbiology consulted covers automatic milk systems and capsule-machine waste; steam wands were not tested.9,10,11

Key studies#

Müller et al., 2015 – Metals from coffee machines and kettles. Filled 8 coffee machines and 11 kettles with food simulants and measured the metals released, before and after descaling. Portafilter machines released the most, and release rose sharply after descaling; one portafilter machine released much less, which the authors read as showing that high release can be avoided.

Food Additives & Contaminants Part A 32(11), 1959–1964 · doi:10.1080/19440049.2015.1086929

Konieczyński et al., 2025 – Elements in café espresso. Measured macro and trace elements by ICP-MS in 104 espressos from Poland and Portugal and in the tap water used. Nickel and lead were far higher in the espresso than in the water; for nickel, the authors state that their method could not tell whether it came from the coffee or the machine.

Foods 14(3), 426 · doi:10.3390/foods14030426

Berg et al., 2000 – Nickel from kettles and coffee machines. Tested electric kettles sold in Denmark under simulated household use, and coffee machines in the same way; the abstract does not name the machine types. Ten of 26 kettles released more than 50 µg/L of nickel, the action threshold, but the coffee machines released no aluminium, lead, chromium or nickel in significant amounts.

Food Additives and Contaminants 17(3), 189–196 · DTU Orbit record

Vilanova et al., 2015 – Bacteria in capsule-machine waste. Sampled the waste reservoirs of capsule machines used for at least a year and followed a new machine over time. All machines held a varied bacterial community; enterobacteria came first and a more stable community formed within a month.

Scientific Reports 5, 17163 · doi:10.1038/srep17163

Contested
  • Nickel and lead from coffee machines: a 2000 Danish study found that coffee machines tested like kettles released no aluminium, lead, chromium or nickel in quantities of any significance7; a 2015 German study found releases of lead, nickel, manganese, chromium and zinc within and partly beyond the Council of Europe release limits in portafilter machines, with release rising sharply after descaling.5 The Danish abstract does not name the machine types tested.
Share or cite this entry

Change log

  • – The BfR rinse advice now reads as a water-only rinse before each preparation in daily use, not only before the first coffee of the day. Reason: The BfR FAQ of 7 May 2025 recommends a rinsing step before the actual preparation of espresso or coffee in daily use; it does not say first of the day.
  • – Zürich 2024 results now state 4 of 15 samples above the Enterobacteriaceae guideline value plus one further case above the aerobic mesophilic guideline; the 'one in four or five' summary was replaced by the counts. Reason: The Kantonales Labor Zürich release reports a further case beyond the four; the total above any guideline is not stated.
  • – Müller et al. 2015: portafilter releases now 'within and partly beyond' the Council of Europe limits (was 'reached or exceeded'); the low-release machine now shows release above the level of concern is 'technically avoidable' (was 'by design'). Reason: Wording of the abstract (read in the science.gov record).
  • – Konieczyński et al. 2025: the statement that the source could not be identified is limited to nickel; for lead the authors' discussed reasons are given. Reason: The paper states the source limitation for nickel only.
  • – Removed an inference that acid descaling acts much less on silica scale (drawn from a statement on rainwater weathering of silicate rock) and an uncited claim that the steam boiler gets more scale because it runs hotter; removed an unsupported citation for coffee residues in the group. Reason: The cited passages do not support these statements.

Sources

  1. A/BWellinger, M.; Smrke, S.; Yeretzian, C. (2018). The SCA Water Quality Handbook, edition 2 (ISBN 978-0-9995807-3-8). Specialty Coffee Association.Quoted passage: a recommendation; Water handbook of the Specialty Coffee Association; scale amounts are calculated with a modified Langelier index, not measured in machines..
  2. BWellinger, M.; Smrke, S.; Yeretzian, C. (2017). Water for Extraction – Composition, Recommendations, and Treatment. In: Folmer, B. (ed.), The Craft and Science of Coffee, Academic Press, ch. 16, pp. 381–398. doi:10.1016/B978-0-12-803520-7.00016-5 · other copiesQuoted passage: a description; Textbook chapter on water for coffee extraction; general account of scale and corrosion in coffee machines, no measurements..
  3. A/BBeeman, D.; Songer, P.; Lingle, T. (ed. Everage, L.) (2011). The Water Quality Handbook (ISBN 978-1-882552-08-5). Specialty Coffee Association of America.Quoted passage: a description; Trade handbook of the SCAA; field observations from brewer service in the United States, not a controlled study..
  4. AGerman Federal Institute for Risk Assessment (BfR) (2025). Some portafilter espresso machines release too much lead (FAQ, updated 7 May 2025). BfR.Quoted passage: a recommendation; National risk-assessment body; eight new household machines tested in 2013 (three portafilter, three pad, two capsule) plus cited state-laboratory surveys from Stuttgart, 2007 and 2011..
  5. AMüller, F. D.; Hackethal, C.; Schmidt, R.; Kappenstein, O.; Pfaff, K.; Luch, A. (2015). Metal release from coffee machines and electric kettles. Food Additives & Contaminants Part A 32(11), 1959–1964. doi:10.1080/19440049.2015.1086929 · other copiesQuoted passage: the authors’ own result; Laboratory study, Germany; 8 coffee machines and 11 electric kettles, release into food simulants before and after decalcification..
  6. AGerman Federal Institute for Risk Assessment (BfR) (2015). Release of metals from coffee machines and electric kettles (BfR Communication No 041/2015, 9 November 2015). BfR.Quoted passage: a description; Short communication announcing the 2015 journal article; states the legal position in the EU..
  7. ABerg, T.; Petersen, A.; Pedersen, G. A.; Petersen, J.; Madsen, C. (2000). The release of nickel and other trace elements from electric kettles and coffee machines. Food Additives and Contaminants 17(3), 189–196.Quoted passage: the authors’ own result; Danish food administration study of kettles and coffee machines on the Danish market under simulated household use; machine types not given in the abstract..
  8. AKonieczyński, P.; Seroczyńska, K.; Wesołowski, M.; Pinto, E.; Couto, C.; Cunha, A.; Azevedo, R.; Almeida, A. (2025). A Comparative Elemental Analysis of Espresso Coffee from Poland and Portugal. Foods 14(3), 426. doi:10.3390/foods14030426 · other copiesQuoted passage: data; Survey, 60 espresso and water pairs from northern Poland and 44 from northern Portugal (cafés, restaurants, petrol stations); ICP-MS; machine source of metals not identified..
  9. AKantonales Laboratorium Basel-Stadt (2024). Milch aus Kaffee-Vollautomaten in Restaurationsbetrieben: Mikrobiologische Qualität (2023). Gesundheitsdepartement des Kantons Basel-Stadt (report, 16 February 2024).Quoted passage: the authors’ own result; Official food control, canton of Basel-Stadt, Switzerland; 19 milk samples from fully automatic coffee machines in 19 restaurants, 2023; judged against guideline values of the Swiss hospitality guideline..
  10. AKantonales Labor Zürich (2025). Mikrobiologische Qualität von Milchgetränken aus Kaffeeautomaten. Kanton Zürich, Gesundheitsdirektion (press release, 7 July 2025).Quoted passage: the authors’ own result; Official food control, canton of Zürich, Switzerland; 15 samples of milk and milk drinks from fully automatic coffee machines with milk function, 2024; judged against guideline values of the Swiss hospitality guideline..
  11. AVilanova, C.; Iglesias, A.; Porcar, M. (2015). The coffee-machine bacteriome: biodiversity and colonisation of the wasted coffee tray leach. Scientific Reports 5, 17163. doi:10.1038/srep17163 · other copiesQuoted passage: the authors’ own result; Capsule machines (Nespresso system) in Valencia, Spain; waste reservoirs of nine machines used for at least one year, plus colonisation of one new machine over time..
  12. A/BCappellini, D. (applicant Carimali S.p.A.) (2020). Device and process for controlling the extraction pressure of coffee in an espresso coffee machine (EP 3 653 086 A1). European Patent Office; priority 18 July 2017, published 20 May 2020.Quoted passage: a description; European patent application for a pump espresso machine; construction only..
  13. A/BTanami, O. (2024). Steam wand treating device (US Patent 11,992,151 B2). United States Patent and Trademark Office; priority 29 October 2019, granted 28 May 2024.Quoted passage: a description; US patent for a steam-wand cleaning device; the hygiene statement is background, not a measurement..

Tier A · authoritative · B · trade and expert · C · general. Supporting quotes for every source are in the data version of this entry. If a publisher’s site refuses access, try the doi link or “other copies”.

The literature for this entry was last searched on 27 Oct 2026. A statement that no source consulted for this entry shows something refers to the sources above; newer or missing work may exist.

Suggest a change

On Espresso machine care. Tell us what you would change and why. A source that supports it helps most.