Why Do Icebergs and Ice Cubes Float When Solids Are Heavier Than Liquids?

Generally, yes. But water is an exception. As trivial as this question may sound, the answer is of life-and-death importance. If ice didn’t float on water, we might not even be here to wonder why.

Let’s see what would happen if ice sank in liquid water. In prehistoric times, whenever the weather got cold enough to freeze the surface of a lake, pond, or river, the ice would immediately have sunk to the bottom. Subsequent warm weather might not have melted it all because it would have been insulated by all the water above it. The next freeze would have deposited another layer of ice on the bottom, and so on.

Before long, much of the water on Earth, except for an equatorial band where it never freezes, would be frozen solid from the bottom up, and there might not be enough time in the warmer seasons to melt it all the way down. The primitive sea creatures from which we evolved might never have had a chance to develop. The world would be relatively barren of life.

The floating of solid water (ice) on liquid water is so familiar to us that we don’t realize that it’s really an unusual phenomenon. When most other liquids freeze, the solid form is denser, heavier, than the liquid, for a given volume. That’s just what we would expect, because in solids the molecules are packed together more tightly than they are in free-flowing liquids, so naturally the solids will be heavier and will sink.

Try it with a liquid that freezes at a conveniently comfortable temperature: paraffin wax.

Add a piece of solid wax to some melted wax, and watch it sink. You’d get the same result with molten versus solid metals, oils, alcohols, and so on. But perform the scientific experiment of placing an ice cube in a glass of water, and you’ll get the opposite result. The ice cube floats.

The reason for water’s contrary behavior lies in the unique way in which water molecules are connected to each other in a piece of ice. They are connected by bridges (hydrogen bonds) between the water molecules. But consider what a bridge does. Brooklynites might say that the Brooklyn Bridge joins Brooklyn to Manhattan, but Manhattanites might insist that it separates Brooklyn from Manhattan. Well, in a sense they’re both right, and that’s literally what hydrogen bonds do to the water molecules in ice: They join the molecules together, but they also hold them a certain distance apart.

So instead of crowding together as tightly as the molecules in other solids do, the water molecules form a sort of open latticework. The molecules are now farther apart in the ice than they were in the liquid, so the ice takes up more space: A given weight of water occupies about 9 percent more space in the form of ice than it does in the form of liquid.

Look carefully at the ice in your freezer’s ice-cube tray. You’ll notice that the cubes have little mountain peaks. In freezing, they had to expand, and being restricted on the sides and bottom, the only direction they could go was up.

If freezing water is confined so that it can’t expand, it will burst the strongest container in the attempt. That’s why a water pipe or an automobile engine can crack when the water inside freezes.

The bridges in ice don’t form all at once at the instant of freezing. As we start cooling water down from room temperature, it gets denser and denser, just like any other liquid, because the molecules are slowing down and don’t require as much elbow room. Most other liquids just keep getting denser and denser until they freeze, and the solid will be densest of all. But not dear old aqua.

Water becomes denser only down to a point. When it is cooled to 39.16 degrees Fahrenheit or 3.98 (let’s call it 4) degrees Celsius, it starts going the other way, getting less dense as you cool it. That’s because some of those bridges are beginning form. Finally, at 32 degrees Fahrenheit (0 degrees Celsius), all the rest of the bridges snap into place, the water freezes into ice, and the density falls suddenly to a value that’s lowest of all. That’s why ice will float on water of any temperature.

The fact that water has a maximum density at about 39 degrees Fahrenheit has further significant consequences for living things. When cold weather cools the surface of a fresh-water lake, the water at the surface becomes denser and sinks. New water takes its place, gets cooled and sinks. This goes on until all the water in the lake has had a chance to be cooled to its maximum possible density, at 39 degrees, and to sink. Only then can die surface water be cooled down those last seven degrees to form ice at 32 degrees Fahrenheit.

By the time a surface layer of ice can form on a lake, then, all the water in the lake will already be at a temperature of 39 degrees. No matter how cold the weather becomes, any water that gets colder than 39 degrees stays at the top (because it’s lighter), and the fish below can never get any colder and freeze. That’s another reason that water’s peculiarities are responsible for permitting the existence of life on earth.

In real-life bodies of fresh water, temperature fluctuations, winds, water currents, and other mixing phenomena will mess up these tidy arguments about neat layers of water temperatures. But all other things being equal (the universal cop-out), the principles we’ve outlined above will prevail.

In the oceans, however, it’s a slightly different ball game. Because of all the salt it contains, seawater doesn’t have a maximum density at 39 degrees. As its temperature goes down, it just keeps getting denser and keeps on sinking, all the way down to its freezing temperature. In order for ice to form on the ocean surface, all the water first has to get down to the freezing point. That happens only during a long, hard winter near the North or South Pole.

Show me a freshwater pond with a layer of ice on the surface and I’ll tell you the temperature of the water at the bottom, without a thermometer.