{ "id": "milk-and-milk-drinks", "url": "https://coffearium.org/entries/milk-and-milk-drinks/", "license": "CC-BY-4.0", "title": "Milk and milk drinks", "summary": "What milk is made of, how a steam wand turns it into foam, why fat and temperature change that foam, how plant-based drinks compare in measurements, and what milk does to coffee in the cup.", "domain": "science", "kind": "Concept", "stage": "brew", "topics": [ "science" ], "topicSections": {}, "lang": "en", "status": "verified", "reviewer": "Coffearium verification pass", "lastVerified": "2026-10-27T00:00:00.000Z", "literatureSearched": "2026-10-27T00:00:00.000Z", "updated": "2026-10-27T00:00:00.000Z", "added": "2026-10-27T00:00:00.000Z", "brief": [ "Steam injection draws air into milk and heats it; studies steamed milk to 65–70 °C.5,6", "Milk proteins adsorb at the bubble surfaces and hold the foam; caseins were enriched in steamed foam relative to whey proteins in one study.4", "Lipids can impair milk foams, particularly where milk fat is partly crystalline, as at about 20 °C; free fatty acids strongly impair foaming.3", "In 27 plant-based drinks sold in Switzerland, foam height at room temperature ranged from 41.5 to 173 mm, against 134.8 mm for milk.1", "In EU law, “milk” means only the normal mammary secretion; purely plant-based products may not be called milk, with listed exceptions.13" ], "quickFacts": [ { "label": "Cow’s milk (two Swiss UHT milks, packaging)", "value": "Fat 3.5–3.7 · protein 3.2–3.5 · carbohydrate 4.9–5.0 g/100 mL" }, { "label": "Steaming end point in studies", "value": "About 65–70 °C" }, { "label": "Main whey proteins begin to denature", "value": "About 70 °C" }, { "label": "Water added by steam", "value": "About 4–6% (v/v) per foaming (one study)" }, { "label": "Protein in plant-based drinks", "value": "0.0–4.0 g/100 mL (27 Swiss products, packaging)" }, { "label": "The word “milk” (EU)", "value": "Reserved for the normal mammary secretion; plant drinks only under listed exceptions" } ], "aka": [ "Milk foam", "Steamed milk", "Milk texturing" ], "linkTerms": [ "milk foam", "steamed milk", "milk drinks", "milk drink", "frothed milk", "plant-based drinks", "plant-based drink", "milk frothing", "milk froth" ], "related": [ "espresso", "espresso-machine", "espresso-machine-care", "coffee-drinks-around-the-world", "serving-temperature", "body-and-mouthfeel", "chlorogenic-acids", "crossmodal-effects", "sensory-science-methods" ], "contested": [ "Fat content and foaming: a 2010 review states that lipids can be detrimental to milk foams, particularly where milk fat is partly crystalline (about 20 °C)3; a 2024 study found that dispersions of skim milk powder gave about 16 times more foam by mechanical mixing than whole milk at the same protein level5; a 2012 Wageningen report found no correlation between fat content and foaming in semi-skimmed and whole milks frothed at about 56 °C.7 Foaming methods, temperatures and milks differ." ], "corrections": [ { "date": "2026-10-27T00:00:00.000Z", "change": "The EU definition of “milk” is now cited to the 2017 Court of Justice judgment that reproduces it; the separate Regulation (EU) No 1308/2013 source was removed and the text says the regulation itself was not consulted.", "reason": "The regulation's wording had been read only in the judgment, so citing EUR-Lex for it overstated what was consulted." }, { "date": "2026-10-27T00:00:00.000Z", "change": "Fat and foam statements narrowed: “in most studies” and “strongest”/“most harmful” replaced by the review's own wording (“can”, “particularly”); the Wageningen comparison of semi-skimmed and whole milk now also gives its conclusion that the whole milk seemed to foam somewhat more stably; that report's frother is no longer called electric.", "reason": "The statements were stronger than the 2010 review abstract and the 2012 report." }, { "date": "2026-10-27T00:00:00.000Z", "change": "Added the 2012 in vitro study's main result that after simulated digestion more chlorogenic acids were bioaccessible with milk than without, depending on fat content.", "reason": "Reporting only the immediate binding gave a one-sided account of the study." }, { "date": "2026-10-27T00:00:00.000Z", "change": "Consumer study: almond and soy are now associated with bitter, sour, grassy and pungent notes (not “beany”); a 2023 study's machine is described as a two-group espresso machine, not a bar machine.", "reason": "Matches the sources' wording." } ], "sources": [ { "id": "s1", "title": "Comparison of physicochemical properties of commercial UHT-treated plant-based beverages and cow’s milk", "authors": "Kopf-Bolanz, K. A.; Villarreal Cruz, M. C.; Walther, B.; Denkel, C.; Guggisberg, D.", "year": 2023, "publisher": "Agrarforschung Schweiz 14, 43–56", "url": "https://www.agrarforschungschweiz.ch/wp-content/uploads/2023/03/043-056_Guggisberg_Food_Milk_EN.pdf", "doi": "10.34776/afs14-43e", "tier": "A", "quote": "Foam heights were quite different, from 41.5mm to 173mm at room temperature", "quotes": [ { "text": "(milk foam height: 134.8mm)", "for": "milk foam height at room temperature" }, { "text": "50.9mm to 203.6mm at 60°C (milk foam height: 179.3mm)", "for": "foam heights at 60 °C" }, { "text": "The protein concentration showed a significant positive correlation with IFH at 60°C.", "for": "protein and initial foam height" }, { "text": "Our main finding was a positive correlation for foaming at 60°C for the PA content and IFH.", "for": "phytic acid and foam height" }, { "text": "Air was added to the liquid via the filter, 0.3L/min for 30s.", "for": "foaming method" }, { "text": "Data are taken from packaging information.", "for": "composition values in Table 2" }, { "text": "some PBBs showed good foamability, comparable to or even better than that of milk.", "for": "some drinks foam as well as milk" } ], "scope": { "claim": "finding", "context": "27 UHT plant-based drinks from Swiss retailers and two UHT whole milks (3.5% fat); lab foam analyser (50 mL, air sparging 0.3 L/min for 30 s) at room temperature and 60 °C; composition from packaging.", "checked": "2026-10-27T00:00:00.000Z" }, "accessed": "2026-10-27T00:00:00.000Z" }, { "id": "s2", "title": "Effects of High-Temperature Milk Processing", "authors": "Deeth, H. C.", "year": 2021, "publisher": "Encyclopedia 1(4), 1312–1321", "url": "https://www.mdpi.com/2673-8392/1/4/98", "doi": "10.3390/encyclopedia1040098", "tier": "A", "tierNote": "Peer-reviewed encyclopedia entry (review reports published with it).", "quote": "β-Lg and α-La begin to denature at ~70 °C", "quotes": [ { "text": "together, they represent ~80% of the whey proteins in milk (β-Lg, ~60%; α-La, ~20%)", "for": "main whey proteins" }, { "text": "Caseins are much more heat-stable than whey proteins", "for": "heat stability of caseins" }, { "text": "One effect of heat on casein is the release of κ-casein from the casein micelle.", "for": "caseins in micelles" }, { "text": "In milk, the reducing sugar is almost always lactose", "for": "lactose" }, { "text": "Lactulose is not present in raw milk and is present in heated milk.", "for": "lactulose from heating" }, { "text": "lactose is epimerized by heat to lactulose", "for": "lactose converted to lactulose" } ], "scope": { "claim": "description", "context": "Review of heat-induced changes in industrial milk processing (pasteurisation, UHT and similar); denaturation onset stated for heat treatment in general, not for steaming.", "checked": "2026-10-27T00:00:00.000Z" }, "accessed": "2026-10-27T00:00:00.000Z" }, { "id": "s3", "title": "Foaming properties of milk: A review of the influence of composition and processing", "authors": "Huppertz, T.", "year": 2010, "publisher": "International Journal of Dairy Technology 63(4), 477–488", "url": "https://onlinelibrary.wiley.com/doi/10.1111/j.1471-0307.2010.00629.x", "doi": "10.1111/j.1471-0307.2010.00629.x", "tier": "A", "quote": "Skim milk foams can be extremely stable, particularly when they are formed at ∼40–50°C.", "quotes": [ { "text": "The presence of lipids can be detrimental to the formation and stability of milk foams, particularly at a temperature where the milk fat is partially crystalline, e.g. 20°C.", "for": "lipids and partly crystalline fat" }, { "text": "The presence of phospholipids, free fatty acids and partial glycerides strongly impairs foaming of milk.", "for": "lipolysis products" } ], "abstractOnly": true, "scope": { "claim": "description", "context": "Review of milk foaming, composition and processing; abstract consulted.", "checked": "2026-10-27T00:00:00.000Z" }, "accessed": "2026-10-27T00:00:00.000Z" }, { "id": "s4", "title": "Ability to re-foam frothed milk at different solid concentrations and their foam structure", "authors": "Ho, T. M.; Lu, Y.-J.; Xiong, X.; Bhandari, B. R.; Bansal, N.", "year": 2024, "publisher": "International Journal of Dairy Technology 77(3), 874–883", "url": "https://helda.helsinki.fi/server/api/core/bitstreams/c104024c-8f61-44c0-9ca4-db61c98ba8c6/content", "doi": "10.1111/1471-0307.13074", "tier": "A", "quote": "foam fractions were more enriched with caseins (i.e. α-CN, β-CN and κ-CN) than whey proteins", "quotes": [ { "text": "Due to lack of tertiary structure, caseins have a great ability to diffuse to the interfacial region", "for": "caseins at the interface" }, { "text": "whey proteins with disulphide bridges and tertiary structure have an ability to form a highly rigid interfacial film", "for": "whey proteins at the interface" }, { "text": "Thus, there is a very high risk of overheating of the foaming milk, leading to the extreme denaturation of proteins", "for": "overheating during steaming" }, { "text": "The steam injection foaming method added a certain amount of condensed steam into milk samples after each foaming, which was about 4–6% (v/v)", "for": "water added by steam" }, { "text": "which can be a result of drainage of liquid film and coalescence of air bubbles", "for": "foam breakdown" } ], "scope": { "claim": "finding", "context": "Reconstituted skim milk powder at 1.5–15% solids, steam injection and mechanical mixing, foamed up to four times; lab, Australia.", "checked": "2026-10-27T00:00:00.000Z" }, "accessed": "2026-10-27T00:00:00.000Z" }, { "id": "s5", "title": "Foaming Properties and Foam Structure of Milk Determined by Its Protein Content and Protein to Fat Ratio", "authors": "Ho, T. M.; Xiong, X.; Bhandari, B. R.; Bansal, N.", "year": 2024, "publisher": "Food and Bioprocess Technology 17(12), 4665–4678", "url": "https://link.springer.com/article/10.1007/s11947-024-03407-y", "doi": "10.1007/s11947-024-03407-y", "tier": "A", "quote": "The prevalent foaming method in coffee shops involves steam injection", "quotes": [ { "text": "creating air bubbles while heating the milk to 65 − 70 °C", "for": "steaming temperature" }, { "text": "with approximately 60% of the initial foam volume collapsing", "for": "collapse of steamed foams" }, { "text": "skim milk powder samples yielded approximately 16 times more foam volume than whole milk powder and whole milk dispersions", "for": "skim versus whole milk, mechanical mixing" }, { "text": "fat can disrupt the formation of intermolecular interactions among proteins", "for": "fat and protein films" }, { "text": "did not affect the foaming properties of whole milk", "for": "added milk powders and whole milk" }, { "text": "their performance depended on the protein-fat ratio", "for": "protein-to-fat ratio" }, { "text": "particularly whole milk powder and whole milk foams, collapsed rapidly within the initial 10 min", "for": "whole milk foams collapse fast" }, { "text": "High levels of products resulting from lipase-catalyzed hydrolysis of milk fat", "for": "lipolysis products" } ], "scope": { "claim": "finding", "context": "Reconstituted skim and whole milk powders, commercial whole milk; steam injection (5 to 65 °C) and mechanical mixing; lab, Australia. The 65–70 °C range describes coffee-shop practice (introduction); the study itself stopped at 65 °C.", "checked": "2026-10-27T00:00:00.000Z" }, "accessed": "2026-10-27T00:00:00.000Z" }, { "id": "s6", "title": "Production of milk foams by steam injection: The effects of steam pressure and nozzle design", "authors": "Jimenez-Junca, C.; Sher, A.; Gumy, J.-C.; Niranjan, K.", "year": 2015, "publisher": "Journal of Food Engineering 166, 247–254", "url": "https://reading-pure-test.eprints-hosting.org/43510/3/1-s2.0-S0260877415002538-main.pdf", "doi": "10.1016/j.jfoodeng.2015.05.035", "tier": "A", "quote": "Milk foams were produced employing different steam pressures (100-280 kPa gauge) and nozzle designs", "quotes": [ { "text": "The steam was injected at a constant flow rate over a period of time which gave a maximum temperature of about 70 C in the milk.", "for": "end temperature" }, { "text": "the air is introduced as a thin layer entrained by the steam jet at its surface.", "for": "air entrainment, plunging jet" }, { "text": "confined-jet and ejector-type nozzles introduce air by the vacuum caused by steam expansion.", "for": "air drawn in by vacuum" }, { "text": "There was a linear increase in bubble size with steam pressure for each nozzle.", "for": "bubble size and pressure" }, { "text": "Thus foams made at 100 kPa drained 92% of liquid at 2 min, whereas foams produced at 280 kPa only drained nearly 80% during the same time.", "for": "drainage at two pressures" }, { "text": "had desired combination of low bubble size, high foam stability and stiffness.", "for": "plunging jet at 180–280 kPa" } ], "scope": { "claim": "finding", "context": "Semi-skimmed pasteurised homogenised milk, 200 mL per foam; three nozzle designs, 100–280 kPa gauge steam; lab, United Kingdom.", "checked": "2026-10-27T00:00:00.000Z" }, "accessed": "2026-10-27T00:00:00.000Z" }, { "id": "s7", "title": "Foaming behaviour of organic and regular milk", "authors": "Pijnenburg, J.; Sala, G.; van Valenberg, H.; Meinders, M. B. J.", "year": 2012, "publisher": "Wageningen UR, project report BO-12.10-004.01-016", "url": "https://edepot.wur.nl/238544", "tier": "A", "tierNote": "University research report (Wageningen UR project report); no stated peer review.", "quote": "No correlation was found between the fat content and foaming properties.", "quotes": [ { "text": "During frothing the milk was automatically heated to about 56°C.", "for": "frothing temperature" }, { "text": "At this temperature there are no fat crystals present in the fat globule that could pierce the fat globule membrane.", "for": "fat crystals" }, { "text": "Therefore, foaming should preferably be performed at temperatures above 40°C.", "for": "foaming above 40 °C" }, { "text": "the addition of linoleic acid has a negative effect on the foaming properties of milk.", "for": "added free fatty acid" }, { "text": "no significant difference in foam ability and foamstability is observed between foam made with", "for": "semi-skimmed and whole milk alike" }, { "text": "Whole A milk seemed to give more stable foams than semi-skimmed A milk.", "for": "conclusions on whole milk" } ], "scope": { "claim": "finding", "context": "Retail organic and regular semi-skimmed and whole milks, Netherlands; 100 mL frothed in an automatic milk frother (not a steam wand) at about 56 °C.", "checked": "2026-10-27T00:00:00.000Z" }, "accessed": "2026-10-27T00:00:00.000Z" }, { "id": "s8", "title": "Effects of different steam injection conditions on cappuccino’s nutritional profile", "authors": "Santini, G.; Klimanova, Y.; Pucciarelli, S.; Polzonetti, V.; Cespi, M.; Perinelli, D. R.; Polidori, P.; Cognigni, L.; Fioretti, L.; Renzi, S.; Vincenzetti, S.", "year": 2023, "publisher": "Food Chemistry 428, 136757", "url": "https://pubblicazioni.unicam.it/retrieve/b0011fd4-c996-4039-b792-e339f34f04f0/cappuccino%202023.pdf", "doi": "10.1016/j.foodchem.2023.136757", "tier": "A", "quote": "A traditional cappuccino is made of 25 mL espresso coffee and 100 mL steam-foamed milk growing to a volume of about 125 mL.", "quotes": [ { "text": "The absence of β-lactoglobulin in UHT milk could be one of the factors", "for": "UHT milk and foam, authors’ interpretation" }, { "text": "that determine the poor stability and consistency of the foams produced with this type of milk.", "for": "UHT milk and foam, authors’ interpretation" }, { "text": "remains constant up to a temperature of about 55 ◦C, and then it begins slowly to decrease", "for": "lactoperoxidase activity" }, { "text": "Steam at a gauge pressure of 110 kPa was injected into milk to reach temperatures ranging from 40 ◦C to 80 ◦C.", "for": "steaming conditions" } ], "scope": { "claim": "finding", "context": "Pasteurised and UHT cow’s milk steamed at 110 kPa gauge to 40–80 °C with the steam wand of a semi-automatic two-group espresso machine; lab, Italy. The cappuccino recipe is the authors’ description, not a specification.", "checked": "2026-10-27T00:00:00.000Z" }, "accessed": "2026-10-27T00:00:00.000Z" }, { "id": "s9", "title": "The Type and Concentration of Milk Increase the in Vitro Bioaccessibility of Coffee Chlorogenic Acids", "authors": "Tagliazucchi, D.; Helal, A.; Verzelloni, E.; Conte, A.", "year": 2012, "publisher": "Journal of Agricultural and Food Chemistry 60(44), 11056–11064", "url": "https://iris.unimore.it/handle/11380/856303", "doi": "10.1021/jf302694a", "tier": "A", "quote": "The addition of milk to coffee caused an immediate decrease in the bioaccessibility due to CGAs binding to proteins.", "quotes": [ { "text": "the increase in bioaccessibility was dependent on fat content", "for": "more bioaccessible after digestion, fat-dependent" }, { "text": "After digestion, 86−94% of bound CGAs remained in the high molecular weight fraction.", "for": "bound CGAs after digestion" }, { "text": "Casein bound 5-caffeoylquinic acid with high affinity", "for": "casein binding" } ], "abstractOnly": true, "scope": { "claim": "finding", "context": "In vitro gastric and pancreatic digestion of coffee with whole, semi-skimmed, skimmed or diluted milk; lab, Italy; abstract consulted.", "checked": "2026-10-27T00:00:00.000Z" }, "accessed": "2026-10-27T00:00:00.000Z" }, { "id": "s10", "title": "Consumer Perception of Milk and Plant-Based Alternatives Added to Coffee", "authors": "Gorman, M.; Knowles, S.; Falkeisen, A.; Barker, S.; Moss, R.; McSweeney, M. B.", "year": 2021, "publisher": "Beverages 7(4), 80", "url": "https://www.mdpi.com/2306-5710/7/4/80", "doi": "10.3390/beverages7040080", "tier": "A", "quote": "Overall, the consumers liking increased when they perceived sweetness in their coffee.", "quotes": [ { "text": "The coffee sample with the milk addition was liked more than the coffee with the oat alternative.", "for": "milk versus oat" }, { "text": "Flavours such as pungent, grassy, earthy, beany, and sour detracted from the acceptability of the coffee samples.", "for": "drivers of disliking" }, { "text": "The almond and soy PBAs were associated with a strong aftertaste, pungent, bitter, sour,", "for": "attributes of almond and soy" } ], "scope": { "claim": "finding", "context": "116 regular coffee drinkers, Nova Scotia, Canada; filter coffee with 1% dairy milk, soy, almond or oat drink (10 g per about 100 g coffee); hedonic scales and CATA.", "checked": "2026-10-27T00:00:00.000Z" }, "accessed": "2026-10-27T00:00:00.000Z" }, { "id": "s11", "title": "Sensory Evaluation – Profiling and Preferences", "authors": "Thomas, E.; Puget, S.; Valentin, D.; Songer, P.", "year": 2017, "publisher": "In: Folmer, B. (ed.), The Craft and Science of Coffee, Academic Press, ch. 18, pp. 419–456", "url": "https://doi.org/10.1016/B978-0-12-803520-7.00018-9", "doi": "10.1016/B978-0-12-803520-7.00018-9", "tier": "B", "quote": "double homogenized milk with lower fat content and lower fat globules led to an increase of the coffee-related attributes, whereas milk with added casein led to a decrease of those same attributes versus whole milk.", "quotes": [ { "text": "only two categories were identified for the latte depending only of the aptitude of the coffee intensity to show or not in the high milk quantity.", "for": "latte and coffee intensity" } ], "scope": { "claim": "cited", "context": "Textbook chapter describing earlier sensory studies (descriptive analysis of white coffee; sorting task on milk dash, cappuccino and latte).", "checked": "2026-10-27T00:00:00.000Z" }, "accessed": "2026-10-27T00:00:00.000Z" }, { "id": "s12", "title": "Barista-Quality Plant-Based Milk for Coffee: A Comprehensive Review of Sensory and Physicochemical Characteristics", "authors": "Gupta, A.; Keast, R.; Liem, D. G.; Jadhav, S. R.; Mahato, D. K.; Gamlath, S.", "year": 2025, "publisher": "Beverages 11(1), 24", "url": "https://www.mdpi.com/2306-5710/11/1/24", "doi": "10.3390/beverages11010024", "tier": "A", "quote": "They found that at high temperatures (above 60 °C), soymilk tends to phase separate into two distinct layers", "quotes": [ { "text": "which is reversible by either cooling the mixture or increasing the soymilk concentration.", "for": "reversibility" }, { "text": "intermediate soymilk concentrations (5–10% w/w) are particularly prone to phase separation", "for": "concentrations prone to separation" }, { "text": "the integration of PMAs—such as almond, soy, oat, and coconut—into coffee beverages has become increasingly prevalent", "for": "plant drinks used in coffee" } ], "scope": { "claim": "cited", "context": "Review; the phase-separation result is from an earlier study of soy drink–coffee mixtures that was not consulted for this entry.", "checked": "2026-10-27T00:00:00.000Z" }, "accessed": "2026-10-27T00:00:00.000Z" }, { "id": "s13", "title": "Judgment of the Court (Seventh Chamber) of 14 June 2017, Verband Sozialer Wettbewerb eV v TofuTown.com GmbH, Case C-422/16", "authors": "Court of Justice of the European Union", "year": 2017, "publisher": "Court of Justice of the European Union (full text as reproduced by IPPT; press release No 63/17)", "url": "https://www.ippt.eu/sites/ippt/files/2017/IPPT20170614_CJEU_VSW_v_TofuTown.pdf", "tier": "A", "tierNote": "The wording of Regulation (EU) No 1308/2013, Annex VII Part III, was read as reproduced in this judgment; the regulation itself was not consulted.", "quote": "must be interpreted as precluding the term ‘milk’ and the designations reserved by that regulation", "quotes": [ { "text": "The term ‘milk’ shall mean exclusively the normal mammary secretion obtained from one or more milkings without either addition thereto or extraction therefrom.", "for": "definition of milk in Regulation 1308/2013, as quoted" }, { "text": "from being used to designate a purely plant based product in marketing or advertising", "for": "purely plant-based products" }, { "text": "even if those terms are expanded upon by clarifying or descriptive terms indicating the plant origin", "for": "descriptive additions" }, { "text": "of the product at issue, unless that product is listed in Annex I to Commission Decision 2010/791/EU", "for": "listed exceptions" } ], "scope": { "claim": "definition", "context": "Preliminary ruling on Regulation (EU) No 1308/2013, Article 78(2) and Annex VII Part III; applies in the EU.", "checked": "2026-10-27T00:00:00.000Z" }, "accessed": "2026-10-27T00:00:00.000Z" } ], "markdown": "## What milk brings to coffee\n\nMilk is mostly water with fat, protein and sugar. Two Swiss whole milks listed 3.5–3.7 g fat, 3.2–3.5 g protein and 4.9–5.0 g carbohydrate per 100 mL.1 These values come from the packaging.1 Its sugar is lactose.2\n\nMilk protein comes in two groups. Caseins form casein micelles.2 Of the whey proteins, two, β-lactoglobulin and α-lactalbumin, make up about 80%.2 Caseins resist heat far better than whey proteins.2 Both groups matter for foam, as the next sections show.\n\nIn the cup, milk proteins bind some coffee compounds, and milk changes how the coffee is liked.9,10 The section “Milk in the cup” covers the evidence. How hot milk drinks are served is covered in serving temperature.\n\n## How steam makes foam\n\nSteam injection is the usual way to foam milk in coffee shops.5 A steam wand on the espresso machine blows steam into cold milk. Two things happen at once: air is drawn in, and the milk heats up.5,6\n\nThe air enters in different ways depending on the nozzle. In one design, the steam jet drags a thin layer of air into the milk at its surface.6 In others, the expanding steam creates a vacuum that sucks air in.6 Studies stopped steaming at about 65–70 °C.5,6 The steam also condenses into the milk. One study estimated that each foaming added about 4–6% water by volume.4\n\nProteins hold the bubbles. They move to the surface between air and liquid and form a film around each bubble.4 Caseins, which lack a rigid folded structure, reach the surface quickly.4 Whey proteins can form a stiff film there.4 In a 2024 study, steamed foam held more caseins than whey proteins, relative to the milk.4\n\n
\n
\n
    \n
  1. Air bubble. Air drawn in by the steam jet.6
    2. \n
  2. Liquid film. Thin layer of milk between two bubbles; liquid drains out of it.4,6
    3. \n
  3. Protein layer. Caseins and whey proteins adsorbed at the bubble surface.4
    4. \n
  4. Fat globule. Fat can disturb the protein film; free fatty acids strongly impair foaming.3,5
    5. \n
  5. Merging bubbles. Bubbles join when the film between them breaks (coalescence).4
    6. \n
  6. Drained milk. Liquid collects below the foam as it drains.6
    7. \n
\n
A milk foam in section, with the protein layer enlarged. Schematic, not to scale.
\n
\n\nFoam does not last. In one study, steamed foams lost about half their volume in 10 min.4 Liquid drains out of the films, and bubbles merge.4 Hygiene of steam wands and milk systems is covered in espresso machine care.\n\n## Fat, heat and foam\n\nFat can work against foam. A 2010 review states that lipids can harm both the formation and the stability of milk foams.3 The effect is particularly marked when milk fat is partly solid, as at about 20 °C.3 Products of fat breakdown, such as free fatty acids, strongly impair foaming.3\n\nTemperature changes the picture. According to the review, skim milk foams can be very stable when formed at about 40–50 °C.3 A 2012 Wageningen report found no link between fat content and foaming at about 56 °C.7 It notes that milk fat holds no crystals at that temperature.7 The studies disagree on how much fat matters; the Deep dive compares them.\n\nToo much heat has costs. A 2024 study warns that steaming can overheat milk and strongly denature its proteins.4 The main whey proteins begin to denature at about 70 °C.2 Steaming to 65–70 °C, as in the studies, stays close to that threshold.2,5,6\n\n## Plant-based drinks\n\nDrinks made from soy, oat, almond and other plants are used in coffee in place of milk.12 In the European Union they may not be sold as “milk”; the Deep dive gives the rule.13\n\nThey differ widely in protein. A Swiss study listed the packaging values of 27 plant-based drinks.1 Soy drinks held 3.0–4.0 g protein per 100 mL, close to milk.1 Oat drinks held 0.3–0.6 g, almond drinks 0.0–1.5 g.1\n\n
\n
\n
Protein in 27 plant-based drinks and two cow’s milks sold in Switzerland, from packaging. Each span covers the products of one base; “<0.5” values are drawn at 0.5.1
\n
\n\nTheir foams differ just as widely. The same study measured foam height after blowing air into each drink.1 At room temperature, plant-based drinks reached 41.5–173 mm; milk reached 134.8 mm.1 Some plant-based drinks foamed as well as milk or better.1 No study consulted for this entry measured steamed foams of named oat, soy and almond drinks side by side.\n\n
\n
\n
Initial foam height of 27 plant-based drinks (range) and of cow’s milk at room temperature and at 60 °C. Foamed by blowing air into the drink in a laboratory analyser.1 Not a steam wand.
\n
\n\n## Milk in the cup\n\nMilk proteins bind coffee compounds. In a 2012 laboratory study, adding milk to coffee at once lowered the free chlorogenic acids, because they bound to milk proteins.9 Casein bound 5-caffeoylquinic acid strongly.9 After simulated digestion, however, more chlorogenic acids were bioaccessible in coffee with milk than without, and the increase depended on the milk’s fat content.9\n\nMilk also changes flavour. In a 2021 study, 116 coffee drinkers tasted filter coffee with dairy milk, soy, almond or oat drink.10 The coffee with milk was liked more than the coffee with oat drink.10 Liking rose when people perceived sweetness; grassy, earthy, beany and sour notes lowered it.10\n\nDrink names vary by place and are not standardised in the sources consulted. One 2023 study describes a traditional cappuccino as 25 mL espresso with 100 mL steam-foamed milk.8 No official cappuccino specification was consulted for this entry. The drinks and their history are covered in coffee drinks around the world and espresso.\n\n\n\n## Proteins at the bubble surface\n\nFoam formation depends on how fast proteins reach and cover a new air–water surface. A 2024 study attributes the speed of caseins to their lack of tertiary structure.4 Whey proteins, with disulphide bridges and a tertiary structure, can form a highly rigid interfacial film.4 In that study, reconstituted skim milk at 1.5–15% solids was steamed up to four times. Foam fractions were more enriched with α-, β- and κ-casein than with whey proteins.4\n\nProtein concentration mattered less than expected. Between 1.5% and 15% solids, foamability and foam stability did not differ significantly.4,5 The same group found that dispersions of skim milk powder gave about 16 times more foam by mechanical mixing than whole milk powder and whole milk at 0.5% protein.5 Adding milk powders or protein concentrate to whole milk did not change its foaming.5 The authors conclude that foaming performance depended on the ratio of protein to fat.5\n\nHeat treatment of the milk may matter too. In a 2023 study, β-lactoglobulin was absent from the UHT (ultra-high-temperature) milk analysed.8 The authors suggest this could be one factor in the poor stability and consistency of the foams made with that milk.8\n\n## Steam injection: pressure and nozzle\n\nA steam wand foams and heats at the same time; the process is not isothermal.6 In a 2015 study, semi-skimmed milk was foamed with steam at 100–280 kPa gauge through three nozzle designs.6 Steam was injected until the milk reached about 70 °C.6 Confined-jet and ejector nozzles draw air in by the vacuum of expanding steam. A plunging jet entrains a thin layer of air at its surface.6\n\nPressure changed the foam, but the effect depended on the nozzle.6 Bubble size rose linearly with steam pressure for every nozzle.6 With the plunging jet, foams made at 100 kPa drained 92% of their liquid in 2 min; foams made at 280 kPa drained nearly 80%.6 The plunging jet at 180–280 kPa gave the best combination of small bubbles, stability and stiffness.6\n\nCondensed steam dilutes the milk. In the 2024 re-foaming study, each steam foaming added about 4–6% (v/v), estimated from volume differences.4 An earlier study cited there gave 10–12% by mass.4 Steamed foams lost 45.50–57.65% of their volume in 10 min across solids levels and repeated foamings.4 In a related study, about 60% of the initial foam volume collapsed within 10 min, particularly fast in whole milk foams.5\n\n## Heat: denaturation and other changes\n\nβ-Lactoglobulin (about 60% of whey proteins) and α-lactalbumin (about 20%) begin to denature at about 70 °C, according to a 2021 review.2 Caseins are much more heat-stable.2 Heat also converts some lactose into lactulose, which is absent from raw milk.2\n\nA 2023 study steamed milk at 110 kPa gauge with the steam wand of a two-group espresso machine to 40–80 °C.8 Activity of the milk enzyme lactoperoxidase stayed constant up to about 55 °C and then slowly fell.8 The 2024 re-foaming study did not test foaming above 65 °C.4 No study consulted for this entry measured how perceived sweetness changes with steaming temperature.\n\n## Lipids and foam stability\n\nAccording to the 2010 review, phospholipids, free fatty acids and partial glycerides strongly impair foaming of milk.3 Free fatty acids and other partial breakdown products form when lipases break down milk fat.5 In the 2012 Wageningen report, added linoleic acid, a free fatty acid, reduced foaming.7 Fat can also disrupt the links between proteins in the film.5\n\nThe state of the fat matters. The review names partly crystalline fat, as at about 20 °C, as particularly detrimental.3 The Wageningen report states that no fat crystals that could pierce the fat globule membrane exist at about 56 °C. It recommends foaming above 40 °C.7 It used an automatic milk frother, not a steam wand.7 At that temperature, semi-skimmed (1.93% fat) and whole milk (4.7%) of one brand showed no significant difference in foamability or foam stability.7 The report’s conclusions add that this whole milk seemed to give somewhat more stable foams.77\n\n## Reference values\n\n| Parameter | Value reported | Conditions | Source |\n|---|---|---|---|\n| Fat · protein · carbohydrate, cow’s milk | 3.5–3.7 · 3.2–3.5 · 4.9–5.0 g/100 mL | Two Swiss UHT whole milks; packaging | Kopf-Bolanz et al.1 |\n| Main whey proteins | β-lactoglobulin ≈60%, α-lactalbumin ≈20% of whey protein | Bovine milk | Deeth2 |\n| Whey protein denaturation begins | ≈70 °C | β-lactoglobulin and α-lactalbumin | Deeth2 |\n| Very stable skim milk foams | Formed at ≈40–50 °C | Review | Huppertz3 |\n| Lipids particularly detrimental | Partly crystalline fat, e.g. 20 °C | Review | Huppertz3 |\n| Steaming end temperature | 65 °C · ≈70 °C | Lab steam wand; 100–280 kPa gauge | Ho et al.5, Jimenez-Junca et al.6 |\n| Steam pressure tested | 100–280 kPa gauge · 110 kPa gauge | Three nozzles · two-group espresso machine | Jimenez-Junca et al.6, Santini et al.8 |\n| Water added by steam | ≈4–6% (v/v) per foaming | Estimated from volume | Ho et al.4 |\n| Foam volume lost in 10 min | 45.50–57.65% | Steamed reconstituted skim milk, 1.5–15% solids | Ho et al.4 |\n| Liquid drained in 2 min | 92% · nearly 80% | Plunging jet at 100 · 280 kPa | Jimenez-Junca et al.6 |\n| Lactoperoxidase activity | Constant up to ≈55 °C | Steamed milk | Santini et al.8 |\n| Foam height, plant-based drinks | 41.5–173 mm (room temperature); 50.9–203.6 mm (60 °C) | 27 UHT drinks; air sparging, 50 mL | Kopf-Bolanz et al.1 |\n| Foam height, milk | 134.8 mm (room temperature); 179.3 mm (60 °C) | Same method | Kopf-Bolanz et al.1 |\n| Casein binding of 5-caffeoylquinic acid | KD 37.9 ± 2.3 µmol/L | In vitro | Tagliazucchi et al.9 |\n\n## Plant-based drinks: what was measured\n\nThe Swiss study foamed 50 mL of each drink by blowing air through a filter at 0.3 L/min for 30 s.1 It read the foam height in a dynamic foam analyser at room temperature and at 60 °C.1 This is a standard laboratory test, not steam injection, so results may differ from a steam wand.\n\nTwo composition factors tracked foam at 60 °C. Protein concentration correlated positively with initial foam height.1 So did the phytic acid content; the authors call this their main finding.1 At 60 °C, foam heights ranged from 50.9 to 203.6 mm, against 179.3 mm for milk.1 The composition values in the study came from packaging, not analysis.1\n\nStability in hot coffee is a separate question. A 2025 review reports an earlier study in which soy drink–coffee mixtures separated into two layers above about 60 °C.12 Cooling or raising the share of soy drink reversed the separation.12 Mixtures with 5–10% (w/w) soy drink were the most prone to separate.12 The original study was not consulted for this entry.\n\nIn the 2021 consumer study, coffee with 1% dairy milk scored 6.0 for overall liking on a 9-point scale. Almond scored 5.7, soy 5.3 and oat 5.2.10 Oat scored significantly lower than milk.10 Milk was associated with sweet and vanilla notes, almond and soy with bitter, sour, grassy and pungent notes.10\n\n## Coffee and milk: chemistry and perception\n\nIn the 2012 in vitro study, milk was added to coffee and the mixture digested with pepsin and pancreatin.9 Milk immediately lowered the chlorogenic acids that passed a 3 kDa filter, because they bound to proteins.9 Casein bound 5-caffeoylquinic acid with a dissociation constant of 37.9 ± 2.3 µmol/L.9 After digestion, 86–94% of the bound chlorogenic acids stayed in the high-molecular-weight fraction.9 Chlorogenic acids are covered in their own entry.\n\nAccording to a sensory science textbook chapter, milk quality changed the sensory profile of coffee with a dash of milk.11 Double-homogenised milk with less fat and smaller fat globules raised the coffee-related attributes; milk with added casein lowered them.11 In a sorting study the chapter describes, latte samples were grouped only by whether the coffee’s intensity showed through the milk.11 No source consulted for this entry measured how milk and foam change body; body and mouthfeel has its own entry.\n\n## The word “milk” in EU law\n\nRegulation (EU) No 1308/2013 reserves the term: “milk” means exclusively the normal mammary secretion obtained from one or more milkings, without addition or extraction. This wording was consulted as quoted by the Court of Justice of the European Union in its 2017 judgment in Case C-422/16 (TofuTown), not in the regulation itself.13 The Court ruled that a purely plant-based product may not be marketed as “milk”, even with words that state its plant origin.13 Exceptions are products listed in Annex I to Commission Decision 2010/791/EU.13 Rules outside the EU were not consulted for this entry.\n\n## Open questions\n\n- **Casein, whey and foam.** Earlier studies reviewed in a 2024 paper disagree on whether casein-rich or whey-rich milk gives more and more stable foam.4\n- **Steam versus laboratory foaming.** Plant-based drinks were compared by air sparging; no study consulted for this entry compared them with a steam wand.1\n- **Temperature and sweetness.** No study consulted for this entry measured how steaming temperature changes perceived sweetness or lactose.\n- **Milk, foam and body.** No study consulted for this entry measured how milk foam changes the perceived body of a milk drink.\n- **Drink definitions.** No official specification for cappuccino, flat white or latte was consulted; the one recipe given here is a study’s description.8\n\n## Key studies\n\n
\n

Jimenez-Junca et al., 2015 – Steam pressure and nozzle design. Foamed semi-skimmed milk with steam at 100–280 kPa through three nozzle types and measured foamability, drainage, bubble size and stiffness. Bubble size rose with pressure for every nozzle; a plunging jet at 180–280 kPa gave small bubbles and stable, stiff foams.

\n

Journal of Food Engineering 166, 247–254 · doi:10.1016/j.jfoodeng.2015.05.035

\n
\n\n
\n

Ho et al., 2024 – Re-foaming steamed milk. Foamed reconstituted skim milk at five solids levels up to four times, by steam and by mixing. Foaming properties did not change with solids or re-foaming, steam added water each time, and the foam was enriched in caseins.

\n

International Journal of Dairy Technology 77(3), 874–883 · doi:10.1111/1471-0307.13074

\n
\n\n
\n

Kopf-Bolanz et al., 2023 – Plant-based drinks and milk. Compared 27 plant-based drinks from Swiss shops with two UHT milks for particle size, colour, viscosity and foam. Foam heights varied widely; some drinks foamed as well as milk, and protein and phytic acid correlated with foam height at 60 °C.

\n

Agrarforschung Schweiz 14, 43–56 · doi:10.34776/afs14-43e

\n
\n\n
\n

Tagliazucchi et al., 2012 – Milk binds coffee chlorogenic acids. Digested coffee with different milks in vitro and measured chlorogenic acids. Milk bound part of them at once through its proteins; casein bound 5-caffeoylquinic acid with high affinity.

\n

Journal of Agricultural and Food Chemistry 60(44), 11056–11064 · doi:10.1021/jf302694a

\n
" }