# Is an ice globe the worst possible way to cool a drink (with ice)?

Given an alcohol solution and a mass of ice in whatever shape you like, is shaping it into a sphere the worst possible way to cool your drink without diluting it?

If the ice starts off at a sub-zero temperature, it is able to cool the drink without diluting it by first warming to 0 °C. After the periphery of the ice hits the melting point, it then waters down the drink in order to absorb more heat. However, if the conduction isn't perfect within the ice, melting could occur before the whole ice cube hits 0 °C, which seems sub-optimal.

I don't know if because it's not just water that's being cooled the thermo changes, or if there's some other important effect I'm neglecting

More important factors are probably the insulating properties of the container and ice, but all things equal, are all those ice-ball claims categorically wrong?

The latent heat of fusion of water is 334 kJ/kg, whereas the specific heat of ice is only about 2 kJ/kg/K, so unless your ice ball is very cold indeed I doubt the warming of the ice to $0^\circ\text{C}$ plays any meaningful role in cooling the drink. Ice is a good thermal insulator, so the melting is pretty much guaranteed to happen as soon as you add the ice, rather than the ice cooling first and then starting to melt, no matter what shape it is.
• @kuzzooroo right, that checks out with the numbers in my post also: $\frac{334\,\text{kJ/kg}}{2\,\text{kJ/kg/K} \,\times\, 18\,\text{K}} \approx 9$. Feb 21, 2018 at 0:50