Kenneth Libbrecht interview: A grand unified theory of snowflakes

SNOWFALL in Pasadena, California, is so rare, it’s almost unheard of. Except, that is, at the California Institute of Technology, where Kenneth Libbrecht can conjure it up using the world’s most sophisticated snowflake-making equipment.

As a physicist, Libbrecht has tackled some fairly epic questions, like the nature of gravitational waves and the internal workings of the sun. But he also has a delightful sideline in the science of snowflakes, which are far more complex and mysterious than you might think. One of the biggest unanswered questions about them is why they appear to come in two distinct types.

Libbrecht went on a 20-year odyssey to solve this mystery. Recently, he published the fruits of that journey in the form of a monograph that runs to more than 500 pages. It contains a kind of grand unified theory of snowflakes, explaining for the first time how and why they grow into the delicate shapes they do.

Joshua Howgego: What got you interested in snowflakes?

Kenneth Libbrecht: One day I was chewing the fat with one of my students and we got talking about how crystals grow and take on shapes. We started thinking about what we could study in this area and I thought: well, water would be cheap and easy. Then I thought: actually, that would be the physics of snowflakes, I wonder how that works? Apropos of nothing – I was just curious – I started reading up on research on snowflakes and I found it really fascinating.

What was the big question about snowflakes?

The standard snowflakes you see in holiday decorations – six-pointed, thin and flat – well, that’s not the way they always form. As well as thin, hexagonal plates, they can grow into tall, hexagonal columns, like the shape of a pencil in miniature. It was discovered in Japan in the 1930s that these two forms of snowflake will form at different temperatures. Plates will form at around -2°C, columns at about -5°C and then plates again at about -15°C. It’s such a crazy pattern that it flops back and forth like that. I really wanted to know why it happens, but it turned out there was no answer – it was a complete mystery.

How did you investigate this conundrum?

I decided the way to answer this question was to systematically grow a lot of snowflakes in different conditions and measure their growth. This was 20 years ago, and for several years I kept hitting problems and made no progress. I eventually figured out the whole experiment has to be enclosed in a box. You add in water vapour, but the conditions, like temperature and pressure, have to be precisely controlled. I have all these little rods going into the box that I use to push things around and turn things on and off. Then I could grow crystals – mostly I grew ones that are smaller than the width of a human hair and then studied them under a microscope. If they get too large, they are too complicated in shape to study easily.

You recently published what has been called a grand unified theory of snowflakes.

People had always thought that if you have a flat surface in a crystal, it always grows in the same way under given conditions. What I found is that in snowflakes, the size of the surface matters quite a lot. If you imagine a hexagonal, plate-like snowflake, it has two wide surfaces and then six much thinner surfaces around the edge. It turns out that those thinner surfaces grow much faster than the broader surfaces, and this creates a runaway effect where you get thinner and thinner plates. The fun thing is that this trend reverses – at certain temperatures, it’s the wider surfaces that grow faster, and so you get column-like crystals.

To understand this fully, you need to dig down into the details about the molecular structure of the ice crystal surface and how it changes with temperature on different surfaces. So far, my model seems to fit all the data, so it is encouraging that at least some of the mystery has been solved.

Does your work tell us anything about what snow would be like on other worlds, such as Saturn’s icy moon Enceladus?

The diffusion of water molecules through the air affects snowflakes’ growth in a way that reinforces the molecular effects to create the thin edges seen in plate-like and hollow-column crystals. When I grow crystals in a vacuum, none of this happens. So, yeah, there would be differences in snow on other planets. If there’s no atmosphere, you are going to get blocky crystals, but at high pressure you would get incredibly thin ones. Of course, the chemistry of the atmosphere could change this, too, in ways that are hard to predict.

You have also grown two identical snowflakes. I thought that was supposed to be impossible.

This started when, early on in my snowflake work, I realised there were no good videos of snowflakes growing. I wanted to make one showing a perfect-looking crystal. Eventually, I worked out how to grow crystals on a fixed support and film them. They looked so much better than crystals that fall from the sky. They were crisper, sharper. Real snowflakes have had kind of a hard life, falling through the atmosphere and banging into other flakes. They have also begun to evaporate, so the edges are always a little rounded.

In the lab, though, I can turn knobs to control the conditions exactly – I can even change things slightly to make the growing crystals branch – and so I can get these designer snowflakes. Then I realised that if you grow two snowflakes next to each other in the box and grow them at the same time, under the same conditions, they come out very similar. Everybody seems to have heard this old adage that no two snowflakes are alike, but then this crazy person is making snowflakes that are alike.

What kinds of questions are still unanswered when it comes to snowflakes?

My model is very complicated. But it makes a lot of predictions and I want to test those out, see what happens and hopefully refine the model. One of the predictions is that interesting things might happen to snowflakes close to their melting point. You might see what’s called pre-melting, where most of the snowflake is a rigid crystal, but on the surface the molecules get disordered. I’ve been trying to explore that. I’m also trying to make larger crystals; I’d like to make the world’s largest snowflake. For no particular reason… there’s just always something new to try.

When you go out in the snow these days, do you see it in a new light?

I grew up in North Dakota, where it gets very cold and we have a lot of snow. I used to see six-pointed star flakes, large ones, but I didn’t know any of this stuff. Now I know a lot better what to look for. I’ll go out and have a magnifier and be looking for different kinds of flakes – like capped columns, for instance, which are like a special hybrid of the plate and column types. I call it snowflake watching. The funny thing is, my wife is a botanist, and when we go out together she’s always looking at the different plants – but they all look like weeds to me.