Field of Science

Cold as Ice

This article in the Atlantic monthly caught my eye, if only because it included an experiment and less because of my refined palate. Wayne Curtis is writing about the unsung hero or villian of mixed drinks: ice.
"I went into the kitchen with another bartender, Stephen Cole, who hunted up a scale and thermometer. He placed the two kinds of ice into separate cups filled with water. We let them sit for 10 minutes. The cheater-ice water proved to be colder (34 degrees compared with 40 degrees), but the ice had lost a full quarter of its weight, compared with just a 14 percent loss in the chunk ice. A cheater-ice cocktail is thus chillier (numbing the taste buds) and more watery (making it flat)."
He describes a bar which stocks eight different types of ice - though the classification system is not quite what a physical chemist might use - or even Kurt Vonnegut. I suspect, however, a serious flaw in the experiment, and therefore in the conclusions drawn about the effect of ice type on a drink.

Take a mixture of ice and water that has been thermally isolated (put in a thermos!) and allow it to come to thermal equilibrium (let it sit until the temperature doesn't change any longer). When the contents of the thermos reach equilibrium, if there are both ice and water present, the temperature is 32 degrees (Fahrenheit). It does not matter how cold the ice was to start, how much water is present, how warm or cold the water was - it will be 32 degrees. Not 40. Not 34.

Also known to those who know how to read a phase diagram, ice at normal pressures will not start to melt until it reaches 32 degrees, and its temperature will not rise above 32 degrees until it has all melted. Curtis' experiment isn't quite as sophisticated as the thermos one I've sketched out, but assuming that the rate of heat loss to the room was small (air - or any gas - isn't a very good thermal conductor, so over the short term this is not a bad assumption), and that the ice and water used were pure, and that a very large amount of water was used relative to the ice - I find it untenable that the "cheater-ice" cocktail is different in temperature than the one made with less porous ice. More watery, yes, colder, no.

Photography by Sue Stafford. Used under Creative Commons license.


  1. I've question sort of related.
    I have noticed that 1% milk seems to freeze at a higher temperature than 2%, whole milk and water.

    Is this a real observation or frustration of finding my milk frozen at dinner time?

    Is its higher freezing temperature than water caused by its calcium?

    And is its lower fat content then other milks causing it to freeze at a higher temperature?

  2. Your suggestion of `a serious flaw' does not seem to be supported by the following paragraphs. Yes, at equilibrium the block of ice will remain at 32F. But this is irrelevant. After 10 minutes, a drink will not typically have reached equilibrium; its temperature is likely to still be higher. This is what Curtis reports: one drink has cooled to 40F, the other 34F. They are continuing to cool, but the one with "cheater-ice" has cooled more.

    And yes, `[ice's] temperature will not rise above 32 degrees until it has all melted', but the drink will be warmer than this.

    If the drink with "cheater-ice" has become waterier than the other drink, then more moles of water must have changed state from solid ice to liquid water. And as the energy required for this change of state is equal to the standard enthalpy of fusion multiplied by the number of moles melting, the "cheater-ice" drink must have been cooled by a greater amount.

  3. The quasi equilibrium state is reached pretty quickly in such small containers. Put a thermometer in a glass with water and ice, and note the temperature every few minutes. Before ten minutes, you will have a straight line at roughly 32 F, this will stay at 32 F until all the ice melts.

    If the ice and water are in thermal contact and in quasi equilibrium state - the water will not be at a different temperature than the ice. That's the point.

    1. I just ran that experiment with 35 gms of ice (crushed on glass A and household cubes in glass B). Both glasses had 175 gms of 72 degree tap water in them. Room temp 72 degrees. Quick stir to each like mixing a drink and crushed ice drink was down to 42 degrees in 1 minute; 40 degrees at 10 minutes with majority of ice gone but not all. Cubed ice 54 degrees at 1 minute and 42 degrees at 10 minutes with a little more than half of the ice dissolved. I believe this supports the theory that smaller ice pieces chill drinks yo a lower temperature in the first minute (which it may be argued is the most important sip of a cocktail) and adds water to the drink at a faster rate. Also,it would have taken more ice than I usually put in a drink to reach 32 degrees within 10 minutes.

  4. Michael --

    Milk isn't a simple solution, but a colloidal suspension, so it's a little tougher to answer this. It's not just the calcium (or probably even principally the calcium) that cause milk to freeze at a different point than water!


Markup Key:
- <b>bold</b> = bold
- <i>italic</i> = italic
- <a href="">FoS</a> = FoS