bulge or explode in the freezer. On the  other hand, when water is compressed,  

it melts. You need a lot of pressure - an ice  cube that's 4 degrees celsius below freezing  

can withstand around 500 times atmospheric  pressure before it'll melt - but it will melt! 

So if when you freeze water, it expands, and  when you compress it, it melts... what happens  

if you try to freeze it while it's compressed?  Like, what if you cool water below 0 degrees  

celsius inside a super strong pressure vessel that  can't bulge or stretch? If the water is liquid,  

it's not under pressure, so below 0 celsius it  should freeze. But if it's frozen, it expands,  

creating pressure, and so it should melt. Which  means it should freeze. Which means it should  

melt. You see the problem. The pressure  that causes melting is being generated  

by expansion that requires being frozen...  it seems like we've arrived at a paradox. 

If by "paradox" you mean "phase diagram"! -  you know, those pictures that tell you whether  

a substance is solid, liquid, or gas - aka, its  phase of matter - for different combinations of  

temperature and pressure. On the phase diagram for  water, you can see here that at normal atmospheric  

pressure, when you cool water down, it goes from  a liquid to a solid, as you'd expect. And while at  

minus 4 degrees celsius and atmospheric pressure  water is frozen solid, when you increase the  

pressure, the solid water goes back to a liquid. So - what if you don't let the water expand when  

it freezes? Well, when water's below 0 degrees  celsius, it wants to freeze - but only some of  

it can, because any bits that freeze, expand,  which pressurizes the container, and eventually  

the pressure builds up enough to keep any more  liquid water from freezing at that temperature!  

On the phase diagram, we follow the line between  liquid and solid in the direction of colder  

temperature & higher pressure. If you cool the  container even more, then slightly more bits of  

water will be able to freeze and expand, until the  pressure again builds up enough to keep any more  

liquid water from freezing. And so on. The colder  you make the container, the higher the percentage  

of ice, and the higher the pressure in the  container. This graph shows the percentage of ice  

vs liquid water when you cool a fixed volume of  water to different temperatures, and I've labeled  

pressures generated at each temperature, too. As for whether the container ever completely  

turns to ice... Well, looking at the phase diagram  again, we see that once the temperature is low  

enough and the pressure high enough, the remaining  liquid water can freeze into a different phase of  

ice, called ice III. And ice III contracts and  becomes denser when it freezes, creating more  

space and allowing the entire container to freeze  solid - though some will still be our normal ice  

(called ice Ih) and some will be ice III. So there is no paradox - the phase of water,  

it turns out, can be non-binary. Wait! Water - water you doing? Don't click  

away - ice suspect you might be interested in this  video's sponsor, Brilliant. Brilliant is a fun,  

interactive science and math learning platform  for curious people young and old, professional  

and beginner. Brilliant is based off the principle  that learning by doing is the fastest path towards  

mastery of a new concept or skill - not just  lectures - and they have interactive courses  

ranging from logic to fundamentals of computer  science to quantum mechanics to infinity and  

beyond (literally - they have courses about  infinity)! To gain a deeper understanding  

of science and mathematics and to sign up for  free, go to Brilliant.org/MinutePhysics. The  

first 200 people will get 20% off an annual  Premium subscription with full access to  

all of Brilliant's courses and puzzles, with  more exclusive content added monthly. Again,  

that's Brilliant.org/MinutePhysics - and  thanks to Brilliant for their support.