Here is a great video of some neat parlor tricks possible with supercooled water, a strange state in which pure water remains liquid after being cooled below zero Centigrade… until something triggers the sudden nucleation of solid ice:
There is a sinister side to nucleation, too. But first, let’s explain what is going on in the video. Water cooled below the freezing point is in a meta-stable state, in the sense that it could minimize its free energy by crystallizing, but is so pure that the system is “stuck” until something gets the freezing started. Of course, once it gets going, the process is self-sustaining and won’t stop until the entire thing is frozen. In the video, supercooled water can use either a seed crystal (regular ice), or just a sharp shock to start freezing. Usually, some dust or other impurity is enough to nucleate freezing, so it takes a lot of preparation to successfully pull of the tricks shown.
This reminded me of the novel Cat’s Cradle, in which Kurt Vonnegut does a fantastic job of explaining how the (fictional, luckily for us) super-stable allotrope of water called ice-nine could be really dangerous.“…Now think about cannonballs on a courthouse lawn or about oranges in a crate again,” he suggested. And he helped me to see that the pattern of the bottom layers of cannonballs or of oranges determined how each subsequent layer would stack and lock. “The bottom layer is the seed of how every cannonball or every orange that comes after is going to behave, even to an infinite number of cannonballs or oranges. Now suppose,” chortled Dr. Breed, enjoying himself, “that there were many possible ways in which water could crystallize, could freeze. Suppose that the sort of ice we skate upon and put into highballs–what we might call ice-one– is only one of several types of ice. Suppose water always froze as ice-one on Earth because it had never had a seed to teach it how to form ice-two, ice- three, ice-four…? And suppose,” he rapped on his desk with his old hand again, “that there were one form, which we will call ice-nine–a crystal as hard as this desk–with a melting point of, let us say, one-hundred degrees Fahrenheit, or, better still, a melting point of one-hundred-and-thirty degrees.”
The problem of course, is the liquid we call “room-temperature water” is supercooled with respect to ice-nine, so it only need a single seed-crystal to take over the system.
In a less apocalyptic application, supersaturated solutions, like sucrose dissolved in water, can be used for more delicious purposes, including making rock candy.
The trick to getting large, delicious crystals of sugar instead of a glob of sticky mess is to start with a solution that is supersaturated. This is easily done by starting with hot water and dumping in a lot of sugar. As the water cools, the ability of the solution to dissolve the sugar decreases, until it starts to “precipitate.” By dropping in a seed crystal to get things starting, sugar molecules keep adding themselves to the existing crystal, making it bigger and bigger until it is good enough for your next party.
In the book Stuff Matters, it is explained that getting chocolate to have the remarkable and tempting properties that it does – staying solid at room temperature, looking smooth and shiny when you pull it from its wrapper, giving off a satisfying “snap” when you break it, and most importantly, melting in you mouth to deploy a payload of sugar and cocoa solids – requires a great deal of engineering to make sure that the chocolate crystallizes correctly. This requires proper nucleation of the right structure.
Mathematically, nucleation happens because the total free energy of a crystal depends on its size. There is a balance between the force of surface tension, which discourages the formation and growth of crystals, and the bond that could be made between the solute molecules, which encourages more precipitation. Very small crystals are unstable because of the large surface tension- they have a lot of surface area and not a lot of volume to have bonds to compensate. In contrast, large crystals have many bonds and a lower surface-area-to-volume ratio. Therefore, new molecules will add themselves to the crystal, making it bigger and even more attractive to other floating molecules. So once a crystal is nucleated, it will grow like crazy. Here, we have a great example of a “bi-stable” system, free-floating molecules vs a crystal, separated by a critical crystal size.
The potentially scary part is that, even without a seed crystal, there is a chance that a cluster will defy the odds and grow large enough to nucleate the entire system to precipitate.
This inherent uncertainty, that the dynamical, nonlinear system can spontaneously and permanently switch states has serious consequences. In some cases, a misfolded protein, called a “prion“, is the actual contagious agent of disease. CJD (the human version of Mad-Cow disease) can start with just one rogue protein that serves as a template to recruit healthy proteins into prions themselves. So even though not a living pathogen, a single prion can multiply like a virus. Based on current knowledge, other protein aggregation diseases, including Alzheimer’s and Amyloidosis might also be thought of a “nucleation diseases,” in the sense that a single amyloid plaque might be formed from normal proteins based just on random fluctuations, with an expected waiting time that depends on the concentration of susceptible proteins. So understanding the physics behind nucleation can help us better control these processes, but not remove all of the randomness.