Why does this ice shatter?

May 21, 2009 5:33 AM

Question for a friend: "When we make ice in ice cube trays w/our local spring water, it doesn't come out in whole cubes like it does when we use tap water. It fractures. Is this because of the high mineral content? What's going on and can you explain why?"
posted by Pressed Rat to science & nature (9 comments total)

Is the spring water the same temperature as your tap?

My guess: one water is significantly colder than the other, and the different freezing rates cause the problem -- freeze the top too fast, it tries to hold back the expansion of the rest of the cube, and shatter when it can't.

Test: Pour a pitcher of tap and a pitcher of spring water. Leave overnight in the same place to equalize temperature, pour each into one tray, place both in the freezer (not on top of each other.) I suspect they'll freeze the same way.
posted by eriko at 5:58 AM on May 21

typing ""why do ice cubes crack" into google i found this page:
http://www.goodhousekeeping.com/home/heloise/kitchen/prevent-ice-cube-jan02 that says that hard water mineral deposits build up stick to the tray and cause the ice to crack.
posted by Osmanthus at 6:01 AM on May 21

Per my interested party, the following additional info:

Both waters used were at room temp.
One (I'm assuming the tap) was filtered to remove chlorine & whatever else those charcoal filters remove (ex- chlorine, lead, copper, mercury, & cadmium).
posted by Pressed Rat at 7:15 AM on May 21

When liquids freeze, impurities tend to stay in the liquid. (The proportion that's excluded is quantified by a "segregation coefficient.") These impurities lower the freezing temperature of the remaining liquid, so the last part to freeze is relatively high in salts and minerals. Metals and semiconductors are purified in part by repeating this process over and over and throwing away the last bit each time.

I suspect that freezing initiates at the bottom of the tray and the surface of the liquid (because the cold air in the freezer is being blown past these two regions). If this is true, then salt- and mineral-rich planar interfaces would be created where these frozen regions meet. The crystalline structure of the ice could be disrupted and especially prone to brittle fraction in these regions. The effect would be diminished when purer water is used.
posted by Mapes at 7:18 AM on May 21 [1 favorite]

Warm the underside of your ice cube tray in running water and all your cubes will pop out uncracked.
posted by caddis at 7:50 AM on May 21

My guess is that tap water has air bubbles from the faucet aerator. The spring water does not. The air bubbles compress with force applie, and allow the ice cube to be more flexible than the brittle spring water cubes.
posted by Midnight Skulker at 7:52 AM on May 21

Spring water is often has a very high mineral content, as has been mentioned above. Mapes almost has this above; as the high-mineral-content water freezes, pure water forms ice, leaving the impure mineral salts in progressively less and less free water. Eventually, these form small semi-frozen inclusions in the ice structure of highly salty water. These inclusions weaken the ice and can cause fracturing as the ice expands and contracts with the temperature variation in the freezer. It's often apparent when remove intact ice cubes which then shatter in contact with warm liquid.

This process is well-understood in the formation of sea ice. The inclusions tend to form as drainage tubes in the ice, known as brine channels (see the "growth of ice" section half-way down). Due to the brine channels, salt-water ice is much weaker and more likely to crack than freshwater ice; you need about double the thickness of sea ice to stand on it safely, for example.
posted by bonehead at 9:43 AM on May 21

I think another effect of dissolved materials on water as it freezes is probably at play here, working along with the concentration of solutes in the unfrozen parts of the cubes.

Namely, the way dissolved minerals reduce the expansion of water as it freezes.

Most liquids contract when they freeze, but pure water expands. The more material dissolved in the water, however, the less expansion until, at 27.4 parts per thousand, the expansion of water as it freezes is abolished altogether (I spent 25 minutes looking for a graph showing the curve of expansion vs. concentration of dissolved materials without realizing there was a good one in an illustration [the 2nd] linked to the excellent page bonehead cites).

Because pure water expands when it freezes, it will liquefy under pressure (this is why ice skates are so frictionless-- they exert enough pressure on the ice to form a very slick layer of water just beneath the blades) and that pressure-induced liquefaction is a critical part of the way most ice trays release their cubes. You twist the tray or pull a lever, and the material of the tray puts pressure on the ice, melting a tiny bit of it which then is no longer stuck to the tray, the pressure points move on, and before you know it, voila!, the cube pops out.

But this won't work with any part of the cube which has a high concentration of solutes due to brine shedding caused by the freezing process: those parts of the cube will stay stuck to the tray and the ice cube will tend to fracture along one of the seams Mapes is talking about.

In the case of a typical plastic ice tray, I'd expect the ice to freeze from the top first because plastic is a pretty good heat insulator, and the minerals to be concentrated toward the bottom of the tray. Then when it was completely frozen and you twisted it to get the cube out, little or no pressure-induced liquefaction would take place at the bottom, the cube would fracture, and the bottom part of the cube would stay stuck in the tray.
posted by jamjam at 3:38 PM on May 21

Thanks to Pressed Rat for posting this question for me. (I am now a member here!) And thanks for all the replies.
posted by Mimidae at 9:17 AM on May 22