In our common understanding, ice needs to be above 0°C to melt. I have now done an experiment: a piece of glass covered with ice on one side and a heating module on the other end. Placing it in a -20°C environment, the heating module starts working and after about 10 minutes the ice melts and disappears. so my question is why ice can melt below 0°C and what is the melt pattern or curve? I'd appreciate it if you could answer.

  • 4
    $\begingroup$ It is not a system in equilibrium, there is a flow of energy, so the temperature is not constant, it becomes hotter the closer you get to the heater. $\endgroup$ Jan 2 at 8:06
  • 2
    $\begingroup$ It is difficult to provide an answer without additional details about the experiment: what is the environment at -20°C? How are temperatures monitored, and where? how thick is the glass? How do you control ice temperature? $\endgroup$ Jan 2 at 9:41
  • 1
    $\begingroup$ It's the temperature of the ice (and water) that matters, not the temperature of the air. $\endgroup$ Jan 2 at 12:31
  • $\begingroup$ Not your question, but pressure can also melt ice. For example, it is possible to ice skate because the pressure under the blade melts some ice. This acts as a lubricant. Or tie weights to a thread. Put the thread on an ice cube with the weights hanging. The thread slowly pulls through the ice. $\endgroup$
    – mmesser314
    Jan 2 at 15:45

2 Answers 2


Thermal radiation can be conducted through glass - you can feel the heat of sunshine through a window. So the thermal radiation from the heater passes through the glass and heats the ice that is next to the glass. If you measure the temperature of the ice that is facing the heater you will find this is greater than $0^o$C, so the ice melts from this side.

  • $\begingroup$ I put a temperature sensor on the top of the glass and it showed that the glass temperature was around -3°C while the ice was melting, and only when the ice was completely melted was the glass temperature greater than 0°C. this result is very surprising. $\endgroup$ Jan 2 at 9:22
  • $\begingroup$ @SmithAlice Was there a line of sight from your temperature sensor to the heater element ? How accurate is your temperature sensor, and did you calibrate it first ? I think you will have to provide more details (a diagram maybe) of your experimental setup if you want more specific comments. Also, there are factors such as impurities that can lower the melting point of ice. $\endgroup$
    – gandalf61
    Jan 2 at 11:55

phase diagram of water

This is a phase diagram for water,if you don't know how to make sense of it just bear in mind that curve BD is called fusion curve.

If you know the pressure of system (the 3D space where process is happening) locate it on vertical axis, draw perpendicular to it, locate the point where it cuts fusion curve, call that point state, from that point draw perpendicular on horizontal axis, the point where perpendicular cuts temperature axis gives you the temperature at which the ice at given pressure will melt. Keep in mind it's layman explanation where some details are intentionally dropped, a good explanation will demand a whole lecture on phase equilibrium.

Now, you may see that above 0.026 atm we can can have ice melt into liquid water, below that if you heat ice it will sublimate ie, directly convert into vapour! You can very well see that 0.026 atm is less than atmospheric pressure (1atm) so we can have ice turn liquid below atmospheric pressure till 0.026 atm! Just temperature of phase transition (melting point) will change (increase).

Also read latent heat of fusion.

The 'enthalpy' of fusion is a latent heat, because, while melting, the heat energy needed to change the substance from solid to liquid at atmospheric pressure is latent heat of fusion, as the temperature remains constant during the process.

heating curve of water This is heating curve for water, as you start heating ice you can see this type of plot between time passed and temperature of system.

The experiment you did was flawed, there are many factors which are responsible for confusing observation.

  • you are measuring temperature of atmosphere whereas you should monitor temperature of ice itself.
  • there should be proper agitation so as to avoid temperature gradient from top to bottom in system.
  • ice is not melting at -20°C, the ice nearest to heat source is hotter than other parts of system it is that which is being turned into liquid. As it melts new layer takes its place and starts melting.
  • thermal conductivity of glass is mere 0.8 W/m K($\frac{1}{256}$ of that of aluminium)while that of ice is twice of it(1.6 W/m K) hence ice is getting heated faster than glass that being reason for:

I put a temperature sensor on the top of the glass and it showed that the glass temperature was around -3°C while the ice was melting, and only when the ice was completely melted was the glass temperature greater than 0°C. this result is very surprising.

  • ice takes heat faster than the glass can heat throughout, but glass near to heat source may be hotter, so at what location you put sensor is important to consider.

Way ahead, Using a heating plate (with temperature sensor) with proper magnetic stirrer and thermometer properly inserted in beaker containing ice and a simple vaccum chamber like this one with pressure gauge is an ideal setup to study 'melt curve' or heating curve and phase transition diagram of water.vaccum chamber


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.