I just want to know for once and for all, exactly why does water expand when it freezes (or, more specifically, when its temperature falls below 4 degrees Celsius)? And does it keep expanding as it gets colder and colder? Is there a limit?
The short of it is that the crystal structure of water (what we call ice) has gaps in it. Basically, water molecules in the liquid state can get closer to one another than they can in the solid state, becuase of the shape of their crystal structure (it’s similar to diamond) and the size of the molecules (tiny).
In liquid form, there are actually short chains of molecules that are crystal-like, and these can wind around one another and slip into the crevices formed by other ones. As a solid, though, the crystals are rigidly locked, making liquid water more dense than solid water.
Water doesn’t keep expanding as it gets colder, but there is a limit of sorts–at 4 degrees C, water is at its most dense. Heat it up, and it gets less dense. Cool it down, and it gets less dense until it freezes, at which point it becomes ice density.
Hope that explains it well!
LL
IIRC it expands because the molecules start to form up all facing the same direction, and the pattern they are starting to form (ice) happens to be slightly less dense than the water is when it is liquid. It’s density is 1.000 g/ml at 4 deg C., and decreases from there in either direction, though it seems to decrease faster as it gets cooler. My CRC handbook only goes to -20 deg C., and it’s density at that point is about the same as water at about 36 deg. C. The only table I have is at a constant pressure of 1.0 atm. too, which is important because the properties change with pressure as well.
I am guessing here, but at some point the pattern it makes has to be fairly complete, and then the ice will start increasing it’s density again. I don’t know what that point is though, and I don’t recall ever needing to know it before. Now ya got me curious though.
well, nope, this doesn’t do it for me. I don’t know WHY the molecules of water should be able to get closer together in their more mobile state, but as they slow down, or crystalize, they are unable to get so close together. Maybe I need some sort of model to help visualize and understand this. Also, as ice gets colder, at what point does it STOP getting less dense and then, does it continue to get MORE dense? For how long? to absolute zero? I’ll keep waiting for an answer. Thanks, so far, though.
Yes, you need a model. A good chemistry text will almost certainly have pictures that’ll do it for you. I can visualize in my mind 'cause I already know what it looks like… but you don’t have that advantage 
The best analogy I can think of is Tinkertoys. When they’re disconnected, they take up a lot less room than after you put them together.
Go to your nearest university library and find a copy of the CRC Handbook of Chemistry and Physics. It will have a table of the density of water at various temperatures.
LL
The table I was using was from my CRC handbook. It stops at -20 C. It’s the 55th editon, 1974-1975, so it’s 25 years old. Maybe a newer version would have it.
If I remember right (it’s been 6 years) isn’t it that water has positive and negative ends?
O -
/ \
H + H +
So that as things cool down, the negative oxygen ends get close to positive hydrogen ends on other molecules.
Like I said, it’s been a while - but if that’s the case, a bit of sketching would provide you with a rough idea of the structure.
I asked the same thing a while back. You might want to look at this thread:
http://boards.straightdope.com/sdmb/showthread.php?threadid=7645
Rysdad - thanks for the thread. Unfortunately, for some reason it’s no longer current, or I can’t get to it, or something. So - I’m still left with my questions and even the teeming millions struggle with a clear explanation. Best so far: the tinker toy analogy. But I’m still not sure why it happens. And - Still doesn’t answer the other question - does water keep expanding as it gets colder until it gets to some temperature and then begin to contract again? What temperature? So - maybe I have to turn this one over to Cece, the great master of all knowledge. Moderator?
The water molecules are asymmetric - one end is positively charged, and the other end negatively charged. Think of them as tiny littme magnets which have N and S poles. If you have a box of those tiny magnets, shake it up a bit and then let it settle, it will form some kind of a pattern - a structure with big gaps in them. However, if you continue to shake it, these structures break up and you are left with just a sea of little magnets juggling around. The volume will decrease, because there are fewer gaps in the continuously shaken box.
On the other hand, many other molecules are symmetric. They are just like beans, not magnets - if you shake it, the volume will incrase because the shaking creates some gaps.
The point everyone is trying to get across, CC, is that ice is a very different solid state than solid states of other substances. Most other solids have their molecules packed tightly together, atomic radius to atomic radius. This is not the case with ice. Water can form special intermolecular bonds called hydrogen bonds. Each molecule can form up to four hydrogen bonds; one for each of the two lone pairs of electrons on the oxygen and one for each of the two hydrogens. Hydrogen bonds have essentially a set length. This length is significantly larger than the molecular radius of water. There is, therefore, a large amount of empty space in between the molecules. When ice is heated there is enough energy to break some of the hydrogen bonds. The liberated molecules tend to fill up the empty spaces.
Once water freezes it will contract as it gets colder. Between 4C and freezing water expands as it gets colder. This is the only time it does that. It expands because the crystaline structure is forming.
Water freezing to ice forms hexagonal rings. The oxygen atoms are at the corners of the hexagon, and the hydrogen atoms are along the edges. The center of the ring is empty, which lowers the density of the crystal. A single water molecule either can’t fit, or it isn’t energetically favorable for it to be there. When ice melts, the rings get broken, and the water molecules can pack more tightly together.
There’s a nice drawing in my college chemestry book, and probably in any other decent introductory college chemistry book as well.
The problem is that the answer to your question is pretty complicated, CC. Since you were unsatisfied with more brief answers, I’ll go into a bit more detail.
There are at least ten different known types of ice (my reference book is from 1986), Ice I[sub]h[/sub] , Ice I[sub]c[/sub] , and Ice II through Ice IX. (OT: ‘Ice Nine’ was a fictional new form of ice featured in a science fiction story. Alas, there is now actually a real Ice IX.) Each has a different crystal structure, or geometric arrangement of atoms in a three-dimentional shape. Ice I[sub]h[/sub] is the kind of ice you’ll get if you cool liquid water at normal atmospheric pressure. Others happen at different conditions; Ice I[sub]c[/sub] is stable at under the same conditions as Ice I[sub]h[/sub] but forms when low pressure water vapor condenses at temperatures of around -120 to -140 degrees C. There exist tables showing at which combinations of temperatures and pressures the different forms of ice are stable. H[sub]2[/sub]O at room temperature and 25 thousand atmospheres of pressure is Ice VII, with a density about half again that of liquid water at room temperature.
It’s not that water can’t form a solid denser than its liquid form, but that there is an alternative low density form which is more stable at atmospheric pressure. Only Ice I[sub]h[/sub] and Ice I[sub]c[/sub] are less dense than liquid water.
The hexagonal analogy others have alluded to is taught in introductory courses, but isn’t quite factual. There are twelve-membered rings in the structure (as ZenBeam referred to), but they aren’t flat and so aren’t really proper hexagons of oxygen. Each oxygen in a I[sub]h[/sub] crystal has four other oxygens nearby, one at 275.2 pm and three at 276.5 pm. The arrangement around each oxygen is nearly tetrahedral, with all O-O-O angles near to 109.5 degrees.
The hydrogen atoms, two per oxygen, are disordered. There is one hydrogen associated with each O-O distance. Each hydrogen has a strong bond (a ‘covalent’ bond) to an oxygen on one side and a weak bond (a ‘hydrogen’ bond) to the oxygen on the other. So, each individual oxygen has two hydrogens covalently bound and two which are weak, longer hydrogen bonds. Both types of Ice I have a disordered crystal structure, which means that for a given oxygen which two of the four tetrahedral directions has a covalent bond and which two have a longer hydrogen bond is random.
The weak hydrogen bonds are long, which keeps the oxygens far apart, which makes the crystal not very dense. There are gaps in the crystal, but they’re three-dimentional, sort of trigonal prisms, not hexagons.
The energy gained by forming these weak hydrogen bonds is greater than the energy cost associated with the lower density of the arrangement. When the pressure gets very high, it’s more efficient to break some or all of the weak bonds in order to fill the space more efficiently, which occurs in the other types of ice. The short answer to why this happens, as with most chamistry questions, is ‘because it’s energetically favorable.’ I hope it makes a little more sense now why it’s energetically favorable in this case.