Why does water freeze at -5C in this experiment?

[natasa_o]

Hi,

I would like to re-define my question from June 5 th. I made an experiment. A bigger surface of rock wool insulation matrix (size 0.4 m x 0.4 m x 0.1 m) was moistened with 45 g of water and sealed with PVC. Then the sealed sample was put into the testing set up. The testing set up consists of two parallel plates, temperature of which can be set. I was changing the temperature of both boudaries (example: T of boundary 1 is 40 C, T of boundary 2 is 20 C for 48 hours, then T of boundary 1 is 20 C, T of boundary 2 is 40 C for another 48 hours) and consequently water evaporated at the boundary with higher temperature and diffused towards the boundary with lower temperature, where the condensation took place (condensation took place also inside the insulation matrix). I measured the temperature ob both boundaries and heat flux at both boundaries. Photo of the speciman and a picture of the testing set up are attached.

When the temperature of the boundary with lower temperature was set to sub 0 C (273K), we found that water freezes at about -5C (268K), but ice melts at 0 C (273 K). During freezing of water at the PVC boundary surface, the temperature of the boundary raises for about 2 K.

Does anybody know why water freezes at -5C? Subcooling? Something else? The sample was moistened with tap water. Is evaporation – diffusion – condensation proces a kind of destilation?

Thanks in advance.

 

[256bits]

When the temperature of the boundary with lower temperature was set to sub 0 C (273K), we found that water freezes at about -5C (268K), but ice melts at 0 C (273 K). During freezing of water at the PVC boundary surface, the temperature of the boundary raises for about 2 K.

Why don't you draw yourselves a spatial temperature profile of your setup and you might see that the conclusion of water freeaing at -5 C might not be all that correct.

Your thermocouple is outside the PVC ( so it looks like ) and would not necessarily measure the temperature of the ice, but that on the PVC/aluminium layer. Even if inside the PVC, the t-couple would be in an air pocket and the temperature of that pocket is what it would measure.

If your hot side is 40 C and the cold side is -20 C, the heat will flow from hot to cold, and your temperature line would slope downwards from left( hot) to right(cold) with several different slopes along the way. The slope through aluminium, being a good conductor, would be more flat, than through the bulk of the insulation. If ice is on the hot side, the melting of the ice, and vaporization, would cause the slope to be more flat than through the bulk of the insulation. The aluminium should record a steeper drop than if melting and vaporization did not occur. At the cold side, the condensation and solidification would also cause the slope to be more flat. as heat is given up at this location you have essentially another heat source which would in essence have an affect of raising the temperature of the aluminium, from that of a control situation where condensation and solidification did not occur. The thermocople would record the temperature between the ice and the alumiinium. The slope through the wool with moisture should be more flat than that without moisture.

Is that your situation?