(First question here on physics, hope it is OK)

Since it is getting cold here in Berlin, Germany and temperatures are falling below the freezing point, we want to make a little ice rink in our back yard (for regular shoes, we just want to slide over the water).

plastic in our backyard

My question is: should I

  1. do it layer by layer: fill the plastic with just a bit of water, wait it for it to get frozen, and fill a bit more, wait for the next freezing and keep on filling and waiting in small amounts until the level is heigh enough?

Or should I

  1. all in once: fill the plastic with enough water and wait (no intermediate steps).

Or perhaps another way?

My goal is to have a solid layer of ice (about 1-2cm - 1/2 inch), not just the top being frozen.

The water from the tap is rather warm, currently about 14°C, probably a bit cooler when I take more than just a (10 liters) bucket.

Edit a day later: I can now report back. Thank you for all the answers. I have filled the plastic with layers of water, waited for it to freeze and then filled another layer and so on.

@Andrew I was thinking about doing the experiment a second area with the other method, but I don't have the same conditions twice so I could not compare each method.

It is hard for me to accept the "right" answer. The "puddle" is frozen, it is a few degrees below 0°C.

Next time I do it, I try without the plastic at the bottom. There are some air bubbles below the plastic and that breaks the ice when sliding. When the (solid) ground is cold enough, the water will freeze anyway.

Edit (2)

The plastic bottom was crap. The ice brake (way too thin) and was not slippery at all. So the next attempt is to remove the plastic and get the ice on the stone floor, layer by layer. (I assume it is getting off-topic now).

water on the terrace

  • $\begingroup$ Use black plastic. $\endgroup$
    – Wouter
    Commented Feb 6, 2021 at 23:29
  • 4
    $\begingroup$ It's more ambitious, but here is a Popular Mechanics article on building a neighbourhood backyard rink in Michigan. Their reference recommended layering the ice making as described below, but it not being a curling rink (where ice properties are MUCH more important) they just filled 'er up and let 'er freeze. Everything worked out just fine. (P.S. I Googled "How to make a backyard rink?" There are lots of articles.) $\endgroup$ Commented Feb 6, 2021 at 23:31
  • 4
    $\begingroup$ Tap water temperature is pretty much irrelevant: With an ambient temperature of -1°C, the temperature differential for 14°C water will be 15 times that of 0°C water, i.e. the iniitial temperature difference will be what goes away fastest. Whether it freezes quickly or slowly depends more on how far below 0°C we are than on initial temperature. $\endgroup$
    – user132372
    Commented Feb 7, 2021 at 1:51
  • 2
    $\begingroup$ Perhaps you can try an experiment on a smaller scale with each of methods 1 and 2. $\endgroup$
    – Andrew
    Commented Feb 7, 2021 at 8:26
  • 5
    $\begingroup$ This seems like a good situation where after your experiments are complete you can provide your own answer as an answer (rather than editing your results into the question). $\endgroup$
    – mattdm
    Commented Feb 7, 2021 at 16:34

4 Answers 4


Water in a fluid phase has density of $1.00~\text{g/cm}^3$, while in a solid form (ice) - $0.92~\text{g/cm}^3$. So while freezing water expands in volume (and thus drops in density), it becomes lighter. That's why we see pieces of ice floating in a river. Due to same water anomalous property, deep lake bottoms may never get completely frozen, because liquid water heat going-up may never reach lake surface and may dissipate in the lower layers of ice. Thanks to that, fish may enjoy living in a winter seasons.

Thus, if your pool is deep enough it may never get frozen fully, or at least will do it slowly if you fill it in one go. A Better tactic is - freezing the ice layer by layer. So that you'll have total control over how thick the ice is. In addition to that, the heat amount contained in a thin layer of fluid water will be lower compared to the case of a fully filled volume, so this heat will escape to surface faster. My advice - fill thin layer of water in the bottom, wait until it freezes to ice. Then fill another layer of water on top of it and repeat the process until whole pit will be complete ice.

  • 8
    $\begingroup$ This feels very off topic for the question. It's asking about a 1/2 inch depth, which would be an extremely shallow pool or pond, more of a puddle if anything $\endgroup$
    – phflack
    Commented Feb 7, 2021 at 4:41
  • 2
    $\begingroup$ @phflack You are right that it is a very thin layer, but nevertheless I was worried about the (theoretical) possibility that the bottom will still be fluid. Thank you for your comment! $\endgroup$
    – topskip
    Commented Feb 7, 2021 at 9:42
  • 3
    $\begingroup$ @Phflack Quite in the contrary. This answer is correct . An ice layer will insulate the remaining water, which is at 4 degrees as this is has maximum density. This slows down ice formation considerably. To grow a reasonable thickness overnight you have to work layer by layer as the experts do. $\endgroup$
    – my2cts
    Commented Feb 7, 2021 at 10:16
  • 2
    $\begingroup$ @my2cts it would be more useful to state the layer thickness, instead of saying just to do it by layers. And while the content may be correct, it's the answer for a different question. Only stating that the sky is blue may be correct, but won't help with a question about why dealing with sorted lists makes loops go faster in computing $\endgroup$
    – phflack
    Commented Feb 7, 2021 at 10:20
  • 1
    $\begingroup$ Great answer and explanation of why to build the rink in layers, rather than just one pool of water. $\endgroup$ Commented Feb 8, 2021 at 5:35

The surface of the water evaporated due to partial pressure. As this transformation from a liquid to a gas state costs energy (the so called latent energy), the remaining water becomes "cold" (decreases its temperature). Thus the more water evaporates per time unit the more ice you obtain. Thus the question is, do you get more gas molecules if the water is in its solid form or in its liquid form.

I'm pretty sure that the vapour pressure is larger in the liquid state as in the solid state. Thus, less water evaporates, once a thin surface of ice is formed which covers the rest of the liquid. Therefore, you should build the ice layer by layer.

Arguments, why I believe that the dominant contribution comes from the latent heat:

  • The heat capacity of the air is smaller than the heat capacity of the ground. Hence, on cold days (or nights) the air is probably colder than the ground.
  • The heat flow is linear in the temperature difference (gradient), $$ \textrm{heat flux} = \mu \cdot \Delta T $$ To freeze the water heat needs to flow from the water to its surrounding (ground and air). The wind constantly replaces the "heated" air above the water by "cold" air. Thus, the temperature gradient is larger between water and air than between water and ground.
  • The "sheet" shown in the above picture is probably a rather good thermal isolator. In addition, there will be a tiny layer of air between the ground and the "sheet". Hence, the coefficient $\mu$ determining the heat transfer between ground and water is probably small.
  • 4
    $\begingroup$ I don't think you can simply ignore the other modes of heat transfer they way you do. $\endgroup$
    – Gert
    Commented Feb 6, 2021 at 15:02
  • $\begingroup$ @Gert: I added some arguments. It would be great, if you could specify the modes of heat transfer you believe are dominant. I believe that heat transfer due to radiation ($\propto T^4$) can be omitted. Which modes are you referring to? $\endgroup$
    – Semoi
    Commented Feb 6, 2021 at 17:54
  • 1
    $\begingroup$ "I'm pretty sure that the vapour pressure is larger in the liquid state as in the solid state." If you're assuming 0°C for both, I would check this claim. $\endgroup$ Commented Feb 6, 2021 at 21:06
  • 3
    $\begingroup$ @Semoi Heat transfer due to radiation is a very large effect at night, if there are no clouds. The reason it doesn't seem very important in everyday life is that you are receiving radiation from every nearby object, so the net amount of radiation is low when all the objects are at similar temperatures. .Deserts (with cloudless skies) have a large difference between day and night temperatures, $\endgroup$
    – alephzero
    Commented Feb 7, 2021 at 4:33
  • $\begingroup$ @alephzero: If we consider the energy loss of the earth / globe, I agree. However, here we are interested in the cooling and freezing of a liquid. Hence, we have to consider the difference of radiation loss between two processes: Putting a large amount of water at once onto the sheet and putting the amount of water "in small steps" on the sheet. What is your argument that the difference is significant? $\endgroup$
    – Semoi
    Commented Feb 7, 2021 at 11:10

For a centimeter or so, I wouldn't try to complicate it. Given cold enough temperatures, that depth can freeze solid in a single night. Also, trying to control the layers thinner than that would be difficult.

If you needed it much thicker, then the layered approach would help. The problem with putting all the water down at once is that basically all of the heat loss is through the upper surface. Once ice forms on the top, it insulates the remaining water and the heat loss slows. By forming the ice in layers, you keep the liquid water as exposed as possible and maximize the heat transfer. If you are diligent about watching when the freezing is complete and put on the next batch of water, it will reduce the total time to freeze.


For fastest freezing make sure the rink is shaded during the day, and exposed to the sky at night.

Optionally you can add environmental snow or ice to the fill water to reduce the amount of cooling (and the amount of tap water) needed allwing you to build thickness more rapidly.

Ice tends to insulate the water below it a bit, so work in layers you'll soon get a feel for how quickly it freezes.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.