What your result means
The two big numbers are the temperatures of two mugs at the moment you drink. Both started as the same tea with the same milk. The only difference is when the milk went in: at the start for one, at drinking time for the other. The difference card shows the gap. A plus sign means milk first is warmer, a minus sign means milk last is warmer.
The two "reaches drinking temp" cards answer the opposite question: how soon you could drink. These two answers point different ways, which is the most useful thing on the page. At the defaults, milk first is warmer after ten minutes, but milk last reaches 60 °C sooner. That is because its cold milk arrives as one drop at the very end. Which one is "better" depends on whether you are waiting a fixed time or waiting to drink.
The last card shows what plain Newton cooling, with a single fixed cooling constant, would say for the same settings. Compare it with the main difference to see how much of the gap comes from physics that Newton's law leaves out.
Why the answer is not what people expect
Most people expect the order not to matter: the tea and the milk carry the same heat whichever way round you combine them. Under textbook Newton cooling with milk at room temperature, that is exactly right. Newton's law says the cup loses heat in proportion to how much warmer it is than the room. Mixing is a weighted average. Both operations are linear, so doing them in either order gives the same answer.
Real cups break that in two ways. First, a cup does not lose heat in proportion to its excess temperature. Evaporation depends on the vapour pressure of water, which rises by about half for every 10 °C near tea temperatures. Radiation depends on the fourth power of absolute temperature. So a cup at 90 °C loses heat much faster than a cup at 80 °C, more than the 10 °C gap alone would suggest. Milk added at the start knocks the tea down to about 84 °C straight away and keeps it out of the most expensive temperature range for the whole wait. Second, more liquid wets more of the mug wall, so the fuller milk-first cup has a little more surface to lose heat from. That works slightly the other way, but it is the smaller effect.
There is a third, quieter effect: fridge milk. In the milk-last mug the milk sits in the fridge at 5 °C until you pour it. In the milk-first mug it has spent the whole wait being warmed by the tea. This happens even under pure Newton cooling, which is why the dashed Newton line shows a small gap at the defaults. Set the milk temperature equal to the room and the Newton gap becomes exactly zero.
How it works
The simulator steps both mugs forward one second at a time using the fourth-order Runge–Kutta method. At every step it works out how many watts the liquid is losing and turns that into a temperature drop, using the heat capacity of the liquid left in the mug. Each step has four parts:
- Convection and conduction from the open top (10 W/m²K), the wetted side wall (8 W/m²K) and the base (5 W/m²K), each multiplied by the gap between tea and room.
- Radiation from the top and the wetted wall: emissivity 0.95 times the Stefan–Boltzmann constant times the difference of the fourth powers of the absolute temperatures.
- Evaporation from the top: saturation vapour pressure from the Magnus formula, turned into vapour density, with the room air at 50 % relative humidity. The evaporation rate comes from the top convection coefficient by the heat and mass transfer analogy, and each gram evaporated takes its latent heat with it. The water lost is removed from the mug, so the milk-last tea is slightly less tea by the time the milk goes in.
- Lid on cuts evaporation to 10 % of the uncovered value and leaves everything else alone.
At 90 °C the default cup loses about 67 W: 13 W by convection, 10 W by radiation and 44 W by evaporation. At 60 °C it loses about 25 W, and evaporation has fallen to 13 W. That steep fall-off is the whole story in one pair of numbers. The dashed reference lines use a single cooling constant, 0.043 per minute. It is fitted so the same uncovered 250 ml mug takes the same time to cool from 90 to 60 °C as the full model, and then held fixed whatever you change.
How to use this simulator
- Set the tea and milk you actually make. A large mug is around 300 to 350 ml and a splash of milk is 15 to 30 ml.
- Set the time you usually wait before the first sip, and the temperature you like to drink at.
- Press Play cooling to watch both mugs cool. Hover or drag across the chart to read every curve at any minute.
- Try the lid, a bigger splash of milk, or milk straight from a warm kitchen, and watch the difference card.
- Copy link saves every setting in the address, and Save picture downloads the chart with your settings and results printed underneath.
Formula
loss: Q = h·A·(T − Tr) + εσA·(TK4 − TrK4) + Lv·km·Atop·(ρv(T) − RH·ρv(Tr))
cooling: dT/dt = −Q / (m·c)
Newton only: T(t) = Tr + (T0 − Tr)·e−kt
Vt, Vm are the tea and milk volumes, treated as grams. Tt, Tm, Tr are the tea, milk and room temperatures in °C, and TK means the same temperature in kelvin. h·A is each surface's convection coefficient times its area. ε is emissivity (0.95) and σ the Stefan–Boltzmann constant. Lv is the latent heat of vaporisation, which falls slightly as temperature rises. km is the evaporation coefficient, ρv the saturated vapour density from the Magnus formula and RH the room's relative humidity. m is the liquid's mass and c its specific heat, 4186 J/kg·K for both tea and milk. k is the fixed Newton constant.
Example calculation
250 ml of tea at 90 °C, 20 ml of milk at 5 °C, a 20 °C room, no lid, drunk after 10 minutes:
After 10 minutes of cooling = 62.8 °C
Milk last: the tea alone after 10 minutes = 64.9 °C, with 6.9 g lost as steam
Add the milk: (243.1 × 64.9 + 20 × 5) ÷ 263.1 = 60.3 °C
Difference: milk first is 2.5 °C warmer
Newton only, same settings: 61.5 °C against 61.1 °C, a 0.4 °C gap
About 2.1 °C of the full model's 2.5 °C gap comes from the nonlinear losses. The fridge-milk effect behind the Newton gap accounts for the remaining 0.4 °C. With a lid the gap falls to 0.9 °C, and with milk at room temperature the Newton gap is exactly zero.
Assumptions & limitations
- Tea and milk have the same specific heat and density as water. Milk's specific heat is a few per cent lower.
- The mug is an 8 cm wide cylinder with walls that store no heat. A real ceramic mug soaks up several degrees from the tea in the first minute. That cools both mugs roughly equally, but it is why real tea cools faster than this model at the start.
- The liquid is perfectly mixed and has one temperature. Real tea is hotter at the bottom and cooler at the surface unless you stir it.
- Convection coefficients, emissivity, 50 % humidity and the lid factor are assumptions chosen to give everyday cooling times. They are not measurements of any particular mug. Draughts, sunlight and a cold worktop all change them.
- The milk-last mug's milk is held at its set temperature right up to pouring, as if it came straight from the fridge.
- The lid is modelled as stopping 90 % of evaporation only. A real lid also traps a layer of warm air and lets condensed steam drip back.
Because of these, treat the size of the gap as an estimate. Which order is warmer is the more robust result, because it comes from the shape of the heat loss curve rather than its exact height.
What to take from it
If you make tea and then get distracted, put the milk in straight away: the tea will be a couple of degrees warmer when you come back to it. If you want to drink as soon as possible, keep the milk back and add it at the end, because the cold milk is the fastest cooling you have. A lid matters far more than the order, worth around 10 °C after ten minutes at the defaults. And none of this touches the separate question of whether milk first changes the taste, which is about chemistry, not heat.
Frequently asked questions
Is tea warmer if you put the milk in first or last?
In this model milk first is warmer when you come to drink it. With the default mug, 250 ml of tea at 90 °C and 20 ml of fridge milk, milk first is at 62.8 °C after ten minutes and milk last is at 60.3 °C. Hot tea loses heat faster than slightly cooler tea, mostly through evaporation, so cooling it a little at the start slows its losses for the whole wait. The gap is a couple of degrees, and a real cup can differ.
Why would the order matter at all?
Under Newton's law of cooling alone, with one fixed cooling constant, it would not matter for milk kept at room temperature: mixing and cooling are both linear, so they can be done in either order. The order matters because real cups do not obey that law exactly. Evaporation and radiation take away proportionally more heat when the liquid is hotter, and a fuller mug wets more of its wall. Fridge milk adds a smaller, separate effect, because milk left in the fridge stays cold until you pour it.
Does a lid change the answer?
A lid keeps milk first ahead, but by less. Evaporation is the most temperature-sensitive way a cup loses heat, and a lid removes most of it. With the defaults the gap shrinks from 2.5 °C to 0.9 °C, and both cups stay far hotter: milk first is at 72.4 °C after ten minutes instead of 62.8 °C.
If I want to drink sooner, should I add the milk first?
No. Holding the milk back gets you to a drinkable temperature sooner, because the cold milk lands as one sudden drop at the end. With the defaults and a 60 °C drinking temperature, milk last reaches it at 10 minutes 12 seconds and milk first at 11 minutes 57 seconds. Milk first wins if you want the tea hotter after a fixed wait; milk last wins if you want it ready sooner.
How accurate is this simulator?
It is a teaching model, not a measurement. The heat transfer coefficients, the mug diameter, the humidity and the lid factor are assumptions, chosen so an open 250 ml mug cools from 90 °C to 60 °C in about 13 minutes. It ignores the heat soaked up by the mug itself, stirring, draughts and steam condensing on a lid. Treat the size of the gap as an estimate. Which order comes out warmer is more robust, because it follows from how heat loss grows with temperature.
Related calculators
Sources and references
The cooling law, the radiation term and the vapour-pressure formula used here come from the following.
- Newton's law of cooling — Wikipedia — the fixed-coefficient law behind the dashed reference lines, and why it is only approximate for natural convection and radiation
- Stefan–Boltzmann law — Wikipedia — the radiation term, with emissivity, and the value of the Stefan–Boltzmann constant
- Vapour pressure of water — Wikipedia — the August–Roche–Magnus formula and coefficients (Alduchov and Eskridge) used for the evaporation term